Close
About
FAQ
Home
Collections
Login
USC Login
Register
0
Selected
Invert selection
Deselect all
Deselect all
Click here to refresh results
Click here to refresh results
USC
/
Digital Library
/
University of Southern California Dissertations and Theses
/
Population genetics and recruitment of the kelp bass, Paralabrax clathratus
(USC Thesis Other)
Population genetics and recruitment of the kelp bass, Paralabrax clathratus
PDF
Download
Share
Open document
Flip pages
Contact Us
Contact Us
Copy asset link
Request this asset
Transcript (if available)
Content
POPULATION GENETICS AND RECRUITMENT OF THE KELP BASS,
PARALABRAX CLATHRATUS
by
Augustus Vogel
A Dissertation Presented to the
FACULTY OF THE GRADUATE SCHOOL
UNIVERSITY OF SOUTHERN CALIFORNIA
In Partial Fulfillment of the
Requirements for the Degree
DOCTOR OF PHILOSOPHY
(BIOLOGY)
December 2006
Copyright 2006 Augustus Vogel
ii
Acknowledgements
This thesis was supported by a large number of individuals, many of which are indicated at
the end of each chapter. I would like to thank Suzanne Edmands (committee chair) and her
lab, including Scott Harrison, Dennis Peterson, Catherine Purcell, Annie Hwang and Sara
Northrup, for their help with the genetic work. Dale Kiefer (committee member) and his lab,
including Vardis Tsontis and Tim Lam, helped with the imagery analysis in Chapter 4.
Otolith work was assisted by the laboratory Oscar Sosa-Nishisaki and Jorge Rosales-Casien,
the laboratory of Gregor Cailliet and Diego Busatto. The rest of the committee, Tony
Michaels, Christine Thacker, Dennis Hedgecock and John Wilson, all helped extensively
with suggestions at multiple stages. Burton Jones, Ivona Cetinic, and Michael Nuemann
answered a lot of questions regarding programming in Matlab®, handling of satellite data,
and interpretation of oceanographic patterns. My graduate career was heavily subsidized by
teaching assistantships, which was supported by the biology office of Bill Trusten, Linda
Bazilian, Glen Smith, Keun Song, and Don Bingham. Other personal financial support came
from Rose Hills Fellowships through the Wrigley Institute for Environmental Studies
(WIES) and research assistantships from Suzanne Edmands. Funding for research included
a fellowship from the National Security Education Program, WIES, QuikScience and the
Packy Offield Foundation.
iii
Table of Contents
Acknowledgements ii
List of Tables iv
List of Figures ix
Abstract xiv
Introduction: 1
References 6
Chapter 1: 9
The influences of sample size and population differentiation
on statistical power and estimation of F
ST
in marine
populations
References 45
Chapter 2: 50
Population genetic structure in Paralabrax clathratus and
Clinocottus analis across potential barriers to gene flow
References 72
Chapter 3: 75
Development and inheritance of molecular markers
in the kelp bass, Paralabrax clathratus
References 89
Chapter 4: 91
Correlating processes associated with ocean triads to
the recruitment of the kelp bass, Paralabrax clathratus
References 128
Chapter 5: 131
Temporal and spatial genetic variance in the kelp bass,
Paralabrax clathratus
References 158
Thesis References: 163
Appendices: 176
Appendix 1 176
Appendix 2 180
iv
List of Tables
Table 1. 17
Simulated values of richness for two mitochondrial control
region (CR) and seven microsatellite (MS) markers. Neutrality
of sequence data was determined in ARLEQUIN (Excoffier
et al., 2005) with Tajima’s D (Tajima, 1989); a Ewens-
Watterson test of homozygosity (F) (Hartl and Clark, 1989)
was performed in “Enumerate” (Slatkin 1996) with
microsatellite data. Allelic richness is given for empirical data
(pre-adjustment) and after simulation with an allele
accumulation curve or Ewens !
k
(Ewens, 1972). Except for
microsatellite Mbo66 (!
k
was set to 1), !
k
was rounded to the
nearest integer in the creation of simulated populations.
Table 2. 21
Power results from sampling of the mitochondrial control
region for differentiated simulated populations. For each drift
goal, 25 population pairs were created, each of which was
sampled 1000 times for each of the sample sizes (for a total
of 25,000 samples).
Table 3. 24
Results from ANOVA tests of the influence of sample size and
population differentiation on the point estimate of power. Data
were normalized with an inverse sine transformation.
Table 4. 25
Results describing the estimation of "
ST
from sampling of
differentiated simulated populations. For each drift goal, 25
population pairs were created, each of which was sampled
1000 times for each sample size (25,000 samples). The 95%
Limits of Data represent the lower and upper "
ST
estimates that
bracket 95% of the data (23,750 samples).
Table 5. 27
Results from 2-way ANOVA analyses on the effects of sample
size and population differentiation on estimation of F
ST
. Tests
included effects on the range, skewness and kurtosis of F
ST
estimates. Data were normalized with an inverse sine
transformation; for "
ST
, the two species were treated as
replicates.
v
Table 6. 32
Sample sizes needed to achieve mean or median estimates of
(1) "
ST
that were at least 90% of the differentiation value for
the two hypothetical drift populations of Paralabrax clathratus
and Clinocottus analis (2) GST and ! that were at least 90%
of the differentiation value. Sample sizes of >100 indicate that
the 90% value was not reached for the sample sizes tested in
this study.
Table 7. 33
Power results from sampling of microsatellite markers for
differentiated simulated populations. For each drift goal, 25
population pairs were created, each of which was sampled
1000 times for each of the sample sizes (for a total of 25,000
samples). Two genetic measures were used: G
ST
and !.
Table 8. 37
Results describing the estimation of G
ST
and ! with sampling
of differentiated simulated populations. For each drift goal, 25
population pairs were created, each of which was sampled
1000 times for each of the sample sizes (25,000 samples). The
95% Limits of Data represent the lower and upper estimates
that bracket 95% of the data (23,750 samples).
Table 9. 54
Record of species genetically surveyed across Punta Eugenia.
Table 10. 57
Sample locations (marked with letters in Figure 1) with sample
size (N), sample dates, count of unique haplotypes in each
population, haplotype diversity (h), sequence diversity (#),
mean pairwise differences (MPD) and Tajima’s neutrality
statistic (D).
Table 11. 64
Results from AMOVA tests with five hypotheses: (1) no
regions, (2) southern region of locations A, B and C and
northern region of all other samples, (3) island region with
populations G, H and I and coastal region with all other
samples, (4) four regions: A, B and C; D and F; G, H and I;
L and M, and (5) a northern region of N and O and a southern
region of all other samples. *p<0.05.
vi
Table 12. 70
Sampling coordinates used for isolation by distance analysis
of data from Fundulus parvipinnis (Bernardi & Talley 2000),
Embiotica jacksoni (Bernardi 2000) and Gillichthys mirabilis
(Huang & Bernardi 2001).
Table 13. 80
Restriction enzymes that recognize polymorphism in the
mitochondrial control region of Paralabrax clathratus.
Polymorphic nucleotide positions correspond to heavy chain
sequences of accession numbers: DQ192295-192399.
Annotated enzymes have the following isoschizomers:
1
AvaII,
SinI;
2
BsiHKAI, HgiAI;
3
Bsp1286I, SduI;
4
Alw44I, SnoI;
5
MvaI;
6
Cfr13I, Sau96I;
7
AquI, BcoI;
8
MaeI, RmaI, XspI;
9
Ce11II, EspI;
10
SecI;
11
BstOI, EcoRII;
12
MflI, XhoII;
13
EcoO109I, PssI;
14
Tsp509I;
15
HindII;
16
MaeII;
17
BfuCI,
Sau3AI, NdeII, DpnI, DpnII;
18
Tru9I;
19
TthHB8I.
**Polymorphism recognized by enzyme SfaNI can be
recognized with AflIII if amplified with primers “K” and
“233cutter.”
Table 14. 84
Comparison between sequence data and RFLP scores for one
hundred and five individuals divided across ten populations:
TR=Tahiguas Reef, NR=Naples Reef, SN=San Nicholas Island,
SC=Santa Catalina Island, CE=San Clemente Island, SQ=San
Quintín, MSD=Morro Santo Domingo, BT=Bahía Tortugas,
BA=Bahía Asunción, AB=Abreojos (see Vogel (in prep) for a
discussion of locations). The RFLP scores were generated with
the restriction enzymes AciI, DdeI, !TaqI, and AflIII. Sequence
results are below the diagonal, RFLP results are above.
*Statistically significant values (p<0.05) are underlined and in
italics.
Table 15. 86
Results for the scoring of eight microsatellites (in boldface, first
column) in the parents and offspring of three crosses of
Paralabrax clathratus. Parental genotypes are indicated above
each marker name. Crosses are identified as male $ female.
Possible allele scores, expected frequencies and observed
frequencies are given for each marker in each cross. Expected
numbers were statistically significantly different for marker
AV17 in cross 6$4; none were statistically different after
Bonferroni correction.
vii
Table 16. 96
Description of areas used in remote sampling of SST data from
1985 to 2005. Letter codes for locations correspond to Figure
11. Standard error for missing pixel percentages is given in
parentheses.
Table 17a. 106
Correlations between age class distributions for samples
collected in 2001 versus 2002/3. Letters represent locations as
correspond to Figure 2. Because (1) sampling in 2002/3 could
collect an extra year or two of younger cohorts than in 2001 and
sample size affected the distribution of rare, older age classes,
correlations were limited to the age classes effectively sampled
by both years. *The correlation between San Clemente Island
samples (D) for example, increases to R
2
=0.6678 (p = 0.0039)
when comparison only includes the main part of the sample from
1988-1997.
Table 17b. 107
Pairwise R
2
values for linear correlations of raw data from the
six adult sampling locations. Letters represent locations as
correspond to Figure 2. Results indicate similarity between the
Coronado Islands (B), San Diego (C), Santa Catalina Island (E)
and Santa Barbara (G). Abreojos (A) and San Clemente Island
(D) appear unique. ***p ! 0.001, **p ! 0.01.
Table 18. 119
Results from correlations between recruitment indices and
indices of the second and third principal components. Imagery
indices were created from yearly summations of loadings that
corresponded to the months of the summer season (May to
September). Putative description of each component is given
before results. ‘Recruitment Data’ is the location from which
samples were taken to create a particular recruitment index.
‘Years’ are the years analyzed and were limited by the
recruitment indices (1995 was removed because of poor
satellite imagery). The R
2
and p-values refer to the linear
correlations between the transformed indices for the principal
component and recruitment. ‘Shape of Relationship’ references
the scatter plot shape of the dependent variable (Eigenvector
loadings index) versus its transformed values which can be seen
in Figures 20-21.
viii
Table 19. 136
Sampling locations names (marked with letters as in Figure 22),
sample sizes (n), number of year classes represented in each
sample, the range of those year classes, sample site
coordinates, and sample dates. Sample SCR98 was collected
by (Findlay and Allen, 2002); sample CPR01 was provided by
J. Caselle and K. Selkoe.
Table 20. 142
Global summary statistics for RFLP analysis of the
mitochondrial control region maker and seven microsatellites.
Table 21. 143
F
IS
scores for locations and location-year classes. Columns
refer to locations in Figure 22. Data grouped by location
(including recruit samples) are in the row “Loc.” Areas of no
data indicate location-year classes that were not collected.
*Significant values of F
IS
are underlined and in italics.
Table 22. 144
Percentage of significant LD scores for locations and
location-year classes. Columns refer to locations in Figure 22.
Data grouped by location (including recruit samples) are in the
row “Loc.” Areas of no data indicate location-year classes that
were not collected. *Percentages of LD that are underlined and
in italics are greater than 0.05 and significant.
Table 23. 145
Pairwise F
ST
estimates for location data. Values above the
diagonal describe microsatellite data and are ! (Weir and
Cockerham, 1984), below describe mitochondrial data and
are "
ST
(Excoffier et al., 1992). *Values underlined and in
italics are significant.
Table 24. 150
Results from testing for significant deficiency of alleles in the
recruit samples. ID’s refer to Table 19 and Figure 22. Sample
‘SCR98’ was not scored for microsatellites. **significant after
sequential Bonferonni correction.
Table 25. 151
!
" F , its 95% confidence interval based on both a %
2
distribution
and a normal distribution and the corresponding values of
!
ˆ
N
e
with adjustments by generation length (G) and a correction
factor for overlapping generations (C). Samples from 1992 to
1997 were analyzed.
ix
List of Figures
Figure 1. 10
Survey of marine population genetics articles in five journals
for conclusions of significant population structure. The
percentage of articles on marine species with planktonic larval
stages that find significant structure is shown. The column
“Total” represents percentages based on summed article counts
and averaged 80.0% over the five year period surveyed.
Figure 2. 28
Plot comparing power sampling of two hypothetical
populations which were drifted to a "
ST
goal of 0.01. Power
estimates represent 25,000 samples taken for each sample size
from Paralabrax clathratus and Clinocottus analis simulated
data.
Figure 3. 29
Plot comparing power sampling of two hypothetical
populations which were drifted to a "
ST
goal of 0.10. Power
estimates represent 25,000 samples taken for each sample size
from Paralabrax clathratus and Clinocottus analis simulated
data.
Figure 4. 30
Plot of sample size versus estimated "
ST
for 25,000 samples
(each sample size) from Paralabrax clathratus mitochondrial
sequence data. Plot includes mean, median and 95% range of
"
ST
estimates for simulated populations that were differentiated
to "
ST
= 0.05.
Figure 5. 35
Plot comparing sample size versus resolution of estimated
power for data from Paralabrax clathratus measured with G
ST
and !. Power estimates from samples populations
differentiated to 0.0005, 0.001, and 0.01 represent G
ST
; a set of
staggered values of 0.001, 0.002 and 0.02 represent !. The
staggered values of ! match the lower values of G
ST
. The
95% confidence intervals have been removed to reduce clutter
in the plot.
Figure 6. 52
Map of sampling locations as described in Table 10.
x
Figure 7. 61
Bayesian consensus phylogeny for 363 bases of the
mitochondrial control region in Paralabrax clathratus with
outgroups of P. maculatofasciatus and P. nebulifer. Posterior
probability values greater than 50% are shown below nodes;
location identifiers are used as described in Table 10 and
Figure 6.
Figure 8. 62
Bayesian consensus phylogeny for 321 bases of the
mitochondrial control region in Clinocottus analis with an
outgroup of Oligocottus maculosus. Posterior probability
values greater than 50% are shown below nodes; location
identifiers are used as described in Table 10 and Figure 6.
Figure 9. 85
Plot of "
ST
scores between ten populations of Paralabrax
clathratus with sequence and RFLP data. Locations are
discussed in (Vogel and Edmands, in prep).
Figure 10. 93
Surface current map along the California coast and Baja
Peninsula as adapted from Durazo and Baumgarter (2002)
and Hickey et al. (2003). Seasonal timing of currents is not
explicitly shown; poleward countercurrents along the near shore
for example, begin in late summer and extend though the winter.
Point Conception (and north) and the tip of the Vizcaíno
Peninsula (Punta Eugenia) are upwelling centers; below them
are the protected cyclonic gyres in the Gulf of Ulloa (GU) and
Southern California Bight (SCB). Both are present during the
late summer recruitment of Parlabrax clathratus but are
replaced by equatorward moving water at other times of the year.
The California Current occurs generally to the east of the SCB
and GU and is consistent in its equatorward direction.
xi
Figure 11. 95
Sample locations of Paralabrax clathratus and areas analyzed
with Pathfinder AVHRR images. Collection locations of fish
include (sample sizes in parentheses): (A) Abreojos, BCS,
Mexico (1002); (B) Coronado Islands, BCN, Mexico (308);
(C) San Diego, CA, USA (532); (D) San Clemente Island, CA,
USA (316); (E) Santa Catalina Island, CA, USA (197);
(F) North Channel Islands, CA, USA (288) and (G) Santa
Barbara, CA, USA (150). Imagery was analyzed for two
locations: around the Vizcaíno Peninsula and Punta Eugenia
(25-28.5N, 112-116.1W) and the Southern California Bight
(31.5-34.5N, 116.6-120.5W).
Figure 12. 100
Plot of Recruitment Indices for the six locations sampled in
this study.
Figure 13. 101
Plot of Recruitment Index for the Northern Channel Islands
from data collected as part of the Kelp Forest Monitoring
Program.
Figure 14. 109
Plots of loading values for principal components 2 and 3 from
images of the Viscaíno Peninsula and Southern California Bight
(SCB). The year 1995 has been removed because a large
number of bad images. The line represents a 7 point moving
average of the data.
Figure 15. 110
Spectral density plots of principal components 2 and 3 from
images of the Viscaíno Peninsula and Southern California Bight
(SCB). Periodicity was found to be seasonal (1 year) for all
components except component 2 of the SCB.
Figure 16. 111
Representative imagery of the Gulf of Ulloa that corresponded
to maximal values of principal components 2 and 3. Positive
values of component 2 were interpreted as retention of a warm
water zone between the point and the coast. Positive values of
component 3 were was interpreted as Ekman upwelling. The
arrow points to Punta Prieta, an apparent southern limit to
upwelling. The northern limit of upwelling is the tip of the
Viscaíno Peninsula, Punta Eugenia.
xii
Figure 17. 112
Derived principal component images for the Gulf of Ulloa.
Components are organized by row, with each image including
average loadings from May 1
st
to September 30
th
. The three
years chosen represent the recruitment success dome in Figure
11; 1994 is found on the left side of the dome because of cooler
than average water along the coast (component two) and high
levels of upwelling (component three). The year 1991 occurs on
the right side of the dome because of excessive warming and
high upwelling. The year 1989 represents a dominant
recruitment year, with moderate arrival of warm southern water
and low levels of upwelling. A combination of components two
and three (14.7% of variance) shows a mean seasonal signal, but
loses the staggered timing of the components.
Figure 18. 116
Representative imagery of the Southern California Bight (SCB)
that corresponded to minimal values of component 2 and
maximal values of principal components 3. Negative values
of component 2 represented upwelling. Positive values of
component 3 described seasonal Ekman upwelling. The arrow
indicates Point Conception, the upwelling boundary between
the SCB and the central Californian coast.
Figure 19. 117
Derived principal component images for the Southern California
Bight (SCB). Components are organized by row, with each
image including average loadings from May 1
st
to September
30
th
. The three years chosen represent the recruitment success
domes in Figure 12; part (a) for the four locations in the SCB
that were compared only to component two and part (b) for the
two locations in the Santa Barbara Channel that were compared
to an index of both components 2 and 3. The years 1992 and
2001 (part a and b, respectively), from the left side of the domes,
had very little upwelling. High levels of upwelling were seen in
1997 and 2004, whereas moderate levels were found in the good
recruitment years of 1989 and 1985. For part b, the high
upwelling year in 1997 also represented high recruitment at the
San Clemente Island sample, in contradiction to the other
locations.
Figure 20. 120
Scatter plots of the Oceanographic Index from imagery of the
Gulf of Ulloa versus their ACE transformed values for
comparisons with the Abreojos recruitment sampling location
(‘A’ in Figure 11).
xiii
Figure 21. 121
Scatter plots of the Oceanographic Index from imagery of the
Southern California Bight versus their ACE transformed values.
Letters indicate locations as corresponding to Figure 11:
(B) Coronado Islands; (C) San Diego; (D) San Clemente Island;
(E) Santa Catalina Island; (F) Northern Channel Islands;
(G) Santa Barbara. Locations B-E are were compared to
principal component 3; locations F and G were a sum of
component 3 and upwelling loadings from component 2
(loadings were transformed to their sign-inverse values).
Figure 22. 137
Sampling locations of Paralabrax clathratus: (AB) Abreojos,
Baja California Sur, MX; (CO) Coronado Islands, Baja
California Norte, MX; (SD) San Diego, California, USA;
(CE) San Clemente Island, California, USA; (SC,
SCR98/SCR04/SCR05 for recruits) Santa Catalina Island,
California, USA; (CPR01 for recruits) Carpenteria, California,
USA; (SA) Santa Barbara, California, USA.
Figure 23. 146
Plot of Component 1 versus Component 2 from Principal
Components Analysis of location and temporal data (plus
recruit samples). Location ID’s correspond to Figure 22.
Eigenvalues for component 1 were 2.953 (25.0% of variance)
for location data and 1.019 (14.4%) for location-year class
data. Eigenvalues for component 2 were 2.145 (18.1% of
variance) for location data and 0.779 (11.0%) for location-year
class data.
Figure 24. 148
Plots of the influence on the "
ST
estimate between Abreojos
(AB) and San Diego (SD) with removal of individual
location-year classes. Estimates are plotted along the x-axis
by the log
10
value of the number of individuals removed. The
dashed line corresponds to the "
ST
estimate for the two samples
without removal of any year class.
xiv
Abstract
Many longstanding questions in marine science center on understanding population
variability. Research efforts that address these issues include analyses of recruitment and
studies of genetic variance. Understanding recruitment variability is important in that it can
change stock abundance and age structure, whereas genetic differences can be reflective of
differential parental contribution, genetic drift and migration. This thesis is a report on
variability in the recruitment and genetic structure in the kelp bass, Paralabrax clathratus,
an important sport fishery in Southern California.
In an analysis of spatial genetic variance, neither Paralabrax clathratus nor woolly sculpin,
Clinocottus analis, were limited by dispersal barriers. This result was used to argue that
Punta Eugenia, reported to be a potential phylogeographic boundary, only limits widely
structured species and does not uniquely impede gene flow.
Analysis of temporal variance showed that Paralabrax clathratus has a small, short-term,
effective population size. Combined with evidence of linkage disequilibrium, unpredictable
patterns of genetic relationship both between locations and location-year classes, and
reduced allelic richness in recruit samples, it was determined that kelp bass exhibit
significant temporal changes with evolutionary and management consequences.
These analyses were supported by a study confirming the utility and Mendelian inheritance
of the genetic markers used. Another study on the influences of sample size and population
differentiation on statistical power and estimation of differentiation also assisted
interpretation. This simulation analysis in more general terms described the need for large
xv
sample sizes and repetitive sampling when studying high gene flow species such as kelp
bass. It further presented some of the mathematical characteristics of methods used to
estimate population differentiation.
The influence of the environment on kelp bass recruitment was analyzed by describing the
correlation between recruitment and triad zones- areas of enrichment and concentration of
nutrients and retention of larvae. Years with median levels of these processes were the most
productive for recruitment, and presented the possibilities of recognizing strong recruitment
years from environmental data and developing generalized explanations of recruitment along
the California and Baja coasts.
1
Thesis Introduction
Many long-standing questions in marine science center on understanding population
variability and its connection to migration and recruitment. These include among others,
variability in recruitment success (Myers, 1998, MacKenzie, 2000), variability in the
geographic patterns of recruitment (Swearer et al., 1999, Bakun, 1996), variability of the
oceanic environment (Espinosa-Carreon et al., 2004, Nezlin and McWilliams, 2003) and
both spatial (Bowen and Avise, 1990) and temporal (Johnson and Black, 1982) genetic
variance. Because of dispersive larval stages and the difficulty in accessing underwater
ecosystems though, we typically have only a snapshot of processes defined by a black box of
events. As such, we often find an obvious pattern (such as variable recruitment) but do not
have immediate, first-hand explanations of the causes. For fisheries species, this is an
especially tenuous situation in that populations can be decimated before we know enough
about the causes of variability to explain how they will rebuild themselves.
This thesis is an effort to study these kinds of population variability in a fisheries species,
Paralabrax clathratus (kelp bass), and in so doing promote a better understanding of the
forces and events that influence population viability. Topics analyzed here include: (1)
geographic genetic variance (Chapters 2 and 5), (2) temporal genetic variance (Chapter 5),
and (3) variance in recruitment success and its connection to environmental variance
(Chapter 4). Chapters 1 and 3 contain methodological analyses that support the other
chapters. Data from this thesis has also been used for an article (Selkoe et al., in review), on
the lack of connection between Mexican and U.S. populations of kelp bass.
2
Topic (1), geographic genetic variance, has been widely studied (Avise, 2000, Rocha-
Olivares and Vetter, 1999, Stepien et al., 2001) and frequently interpreted as indicative of
barriers to gene flow (Bernardi et al., 2003). Evidence of structure is a rationale to
individually manage fisheries stocks or at least be cognizant of directions of migration. In
the marine environment though, most studies find low levels of structure (Ward et al., 1994),
making stock-definition arguments more problematic (Waples, 1998).
One direction to which research has turned in the study low geographic structure is topic (2):
chaotic patchiness (Johnson and Black, 1982, Johnson and Black, 1984) and temporal
genetic variance (Watts et al., 1990, Kordos and Burton, 1993, Hedgecock, 1994b, Edmands
et al., 1996, Moberg and Burton, 2000, Pujolar et al., 2006, Gilbert-Horvath et al., 2006).
Chaotic patchiness, or unpredictable and ephemeral geographic structure that can include
larger genetic differences between proximal populations than distant ones, is theorized to be
caused by temporal mechanisms (reviewed by (Larson and Julian, 1999). These include
arrival of recruits from differentiated populations (Kordos and Burton, 1993), pre or post
settlement selection (Johnson and Black, 1984) or genetic drift from variability in
reproductive success and a small effective number of parents (Hedgecock, 1994a). This
process has management implications, including the need for bet-hedging reserves (Larson
and Julian, 1999) and the recognition that recruitment may have unpredictable aspects.
In concordance with this idea of unpredictability, recruitment studies using non-genetic
techniques have had limited success at finding sustainable correlations between recruitment
variability and environmental variables (Myers, 1998). Although environmental forcers may
not be tightly enough coupled to recruitment (Myers, 1998), an alternative argument poses
3
that synergistic explanations may be more appropriate (Cole, 1999). One synergistic
explanation, covered in Chapter 4 as Topic (3), is the Triad Hypothesis (Bakun, 1996). This
idea posits that in eastern boundary currents (such as those along the California coast), three
processes may be important for successful recruitment. They include: 1) enrichment
(typically from upwelling), 2) concentration of nutrients (through formation of eddies and
fronts) and 3) retention of larvae (also via eddies and fronts). Besides the Mediterranean
(Agostini and Bakun, 2002) and the Benguela Current (Cole, 1999), this pattern has been
demonstrated to occur in the southern California Bight (Cury et al., 1995) and south of Punta
Eugenia along Baja California, Mexico (Bakun, 1996).
The model species for this thesis, Paralabrax clathratus, lives in these two putative triad
areas and could be tested for all three topics. Kelp bass are a temperate Serranid found
between the Columbia river, Oregon, USA and Magdalena Bay, Baja California, Mexico
(Eschmeyer et al., 1983). Kelp bass are an important sport resource: they are part of a $300
million near-shore sport fishery economy (Thomson, 2001) and rank among the top three
most abundantly caught species in the southern California region (Love, 1996). A
substantial number of projects have been performed to understand better the characteristics
and lifecycle of this species; work has been done on their reproductive ability and behavior
(Oda et al., 1993, Erisman and Allen, 2006), larval stages (Love et al., 1996, Cordes and
Allen, 1997, Findlay and Allen, 2002) and movement patterns (Lowe and Topping, 2003).
Like many marine species though, the life cycle of kelp bass is not well enough understood
to permit implementation of complex management strategies. As of 2006, a catch limit of 10
fish/day/person and a size minimum of 12 inches are the limits of intervention. Recruitment,
4
the process of larvae transforming into the adult morph and leaving the plankton, has been
connected to kelp density (Carr, 1994) and intertidal bores (Findlay and Allen, 2002), but
management cannot predict the recruitment success of any particular season. Several small
studies have also been performed on the population genetic structure of the species
(Beckwitt, 1983, Phalen, 1999, Grothues, 1994), but more complex studies such as the
genetic make-up of a recruitment event are only starting to be described (Selkoe et al., in
prep).
Despite the diversity of previous research then, there remained a need for improved
understanding of the species, especially with regards to the three topics presented previously.
These three questions also offered the opportunity of ready-made hypotheses. The
importance of hypothesis testing is addressed, at least for genetics studies, in Chapter 1; it is
an extension of the discussion presented by (Waples, 1998). Chapter 1 focuses on the
influence of sample size and “true” population differentiation on sampling power and
estimation of F
ST
. While both of these factors have recognized, generalized qualitative
influences, Chapter 1 is an effort to describe more quantitatively these effects. This type of
understanding is especially important for high gene flow species (Waples, 1998), such as
Paralabrax clathratus.
One last consideration and the topic of Chapter 3, revolves around methodological issues in
a genetic study. The analyses performed in Chapter 5 could not have been accomplished
without a series of usable genetic markers. Two types of markers were developed for this
project: first, a Restriction Fragment Length Polymorphism (RFLP) assay for the
mitochondrial control region and second, a series of seven microsatellites to analyze the
5
nuclear genome. The use of RFLPs provides a more cost-effect method for analyzing a
marker than sequencing. As compared to sequencing as well, it was shown that equivalent
patterns of population structure could still be found with the method. Chapter 3 also
describes a series of crosses to confirm Mendelian inheritance and linkage equilibrium of the
microsatellite markers. This is a step that is frequently difficult to perform with marine
species because of the need to conduct controlled crosses (Takagi et al., 2003), but is an
important test of the utility of a set of markers. The microsatellites were previously
published as part of a collaboration with K. Selkoe (Selkoe et al., 2005).
6
Introduction References
AGOSTINI, V. & BAKUN, A. (2002) 'Ocean triads' in the Mediterranean Sea: physical
mechanisms potentially structuring reproductive habitat suitability (with example
application to European anchovy, Engraulis encrasicolus). Fisheries Oceanography,
11, 129-142.
AVISE, J. (2000) Phylogeography: the history and formation of species, Cambridge, MA,
Harvard University Press.
BAKUN, A. (1996) Patterns in the Ocean, Ocean Processes and Marine Population
Dynamics, La Paz, BCS, Mexico, California Sea Grant/CIB.
BECKWITT, R. (1983) Genetic Structure of Genyonemus lineatus, Seriphus politus
(Sciaenidae) and Paralabrax clathratus (Serranidae) in Southern California. Copeia,
3, 691-696.
BERNARDI, G., FINDLEY, L. & ROCHA-OLIVARES, A. (2003) Vicariance and dispersal
across Baja California in disjunct marine fish populations. Evolution, 57, 1599-1609.
BOWEN, B. W. & AVISE, J. C. (1990) Genetic structure of Atlantic and Gulf of Mexico
populations of sea bass, menhaden, and sturgeon: influences of zoogeographic
factors and life-history patterns. Marine Biology, 107, 371-381.
CARR, M. (1994) Effects of macroalgal dynamics on recruitment of a temperate reef fish.
Ecology, 75, 1320-1333.
COLE, J. (1999) Environmental conditions, satellite imagery, and clupeoid recruitment in
the northern Benguela upwelling system. Fisheries Oceanography, 8, 25-38.
CORDES, J. F. & ALLEN, L. G. (1997) Estimates of age, growth, and settlement from
otoliths of young-of-the-year Kelp Bass (Paralabrax clathratus). Bulletin of the
Southern California Academy of Sciences, 96, 43-60.
CURY, P., ROY, C., MENDESSOHN, R., BAKUN, A., HUSBY, D. & PARRISH, R.
(1995) Moderate is better: exploring nonlinear climatic effects on the Californian
northern anchovy (Engraulis mordax). Canadian Journal of Fisheries and Aquatic
Science, 121, 417-424.
EDMANDS, S., MOBERG, P. & BURTON, R. (1996) Allozyme and mitochondrial DNA
evidence of population subdivision in the purple sea urchin Strongylocentrotus
purpuratus. Marine Biology, 126, 443-450.
ERISMAN, B. & ALLEN, L. (2006) Reproductive behaviour of a temperate serranid fish,
Paralabrax clathratus (Girard), from Santa Catalina Island, California, USA.
Journal of Fish Biology, 68, 157-184.
ESCHMEYER, W. N., HERALD, E. S. & HAMMANN, H. (1983) A Field Guide to Pacific
Coast Fishes, New York, New York, Houghton Mifflin Company.
ESPINOSA-CARREON, T., STRUB, P., BEIER, E., OCAMPO-TORRES, F. &
GAXIOLA-CASTRO, G. (2004) Seasonal and interannual variability of satellite-
derived chlorophyll piment, surface height, and temperature off Baja California.
Journal of Geophysical Research-Oceans, 109, C03039,
doi:10.1029/2003JC002105.
FINDLAY, A. & ALLEN, L. G. (2002) Temporal patterns of settlement in the temperate
reef fish Paralabrax clathratus. Marine Ecology Progress Series, 238, 237-248.
GILBERT-HORVATH, E., LARSON, R. & GARZA, J. (2006) Temporal recruitment
patterns and gene flow in kelp rockfish (Sebastes atrovirens). Molecular Ecology,
doi: 10.1111/j.1365-294X.2006.03033.x.
7
GROTHUES, T. M. (1994) An investigation into the phenetic structure and larval dispersal
patterns of the kelp bass, Paralabrax clathratus. Marine Biology. Los Angeles,
California State University, Northridge.
HEDGECOCK, D. (1994a) Does variance in reproductive success limit effective population
sizes of marine organisms? IN BEAUMONT, A. (Ed.) Genetics and Evolution of
Aquatic Organisms. London, Chapman and Hall.
HEDGECOCK, D. (1994b) Temporal and spatial genetic structure of marine animal
populations in the California current. California Cooperative Fisheries
Investigations, Reports, 35, 73-81.
JOHNSON, M. & BLACK, R. (1984) The Wahlund effect and the geographical scale of
variation in the intertidal limpet Siphonaria sp. Marine Biology, 79, 295-302.
JOHNSON, M. S. & BLACK, R. (1982) Chaotic genetic patchiness in an intertidal limpet,
Siphonia sp. Marine Biology, 70, 157-164.
KORDOS, L. & BURTON, R. (1993) Genetic differentiation of Texas Gulf Coast
populations of the blue crab Callinectes sapidus. Marine Biology, 117, 227-233.
LARSON, R. J. & JULIAN, R. M. (1999) Spatial and temporal genetic patchiness in marine
populations and their implications for fisheries management. California Cooperative
Fisheries Investigations, Reports, 40, 94-99.
LOVE, M. (1996) Probably More than You Want to Know About the Fishes of the Pacific
Coast, Santa Barbara, CA, Really Big Press.
LOVE, M. S., BROOKS, A., BUSATTO, D., STEPHENS, J. & GREGORY, P. A. (1996)
Aspects of the life histories of the kelp bass, Paralabrax clathratus, and the barred
sand bass, P. nebulifer, from the southern California Bight. Fishery Bulletin, 94,
472-481.
LOWE, C. & TOPPING, D. (2003) Movement patterns, home range, and habitat utilization
of adult kelp bass Paralabrax clathratus in a temperate no-take marine reserve.
Marine Ecology Progress Series, 256, 205-216.
MACKENZIE, B. (2000) Turbulence, larval fish ecology and fisheries recruitment: a review
of field studies. Oceanologica Acta, 23, 357-375.
MOBERG, P. & BURTON, R. (2000) Genetic heterogeneity among adult and recruit red sea
urchins, Strongylocentrotus franciscanus. Marine Biology, 136, 773-784.
MYERS, R. (1998) When do environmental-recruitment correlations work? Reviews in Fish
Biology and Fisheries, 8, 285-305.
NEZLIN, N. & MCWILLIAMS, J. (2003) Satellite data, Empirical Orthogonal Functions,
and the 1997-1998 El Niño off California. Remote Sensing of Environment, 84, 234-
254.
ODA, D., LAVENBERG, R. & ROUNDS, J. (1993) Reproductive biology of 3 California
species of Paralabrax (Pisces, Serranidae). California Cooperative Oceanic
Fisheries Investigations Reports, 34, 122-132.
PHALEN, C. M. (1999) Genetic variation among kelp bass (Paralabrax clathratus) from
seven locations throughout their natural range. Marine Biology. Los Angeles,
California State University, Northridge.
PUJOLAR, J., MAES, G. & VOLCKAERT, F. (2006) Genetic patchiness among recruits in
the European eel Anguilla anguilla. Marine Ecology Progress Series, 307, 209-217.
ROCHA-OLIVARES, A. & VETTER, R. D. (1999) Effects of oceanographic circulation on
the gene flow, genetic structure, and phyogeography of the rosethorn rockfish
(Sebastes helvomaculatus). Canadian Journal of Fisheries and Aquatic Science, 56,
803-813.
8
SELKOE, K., GAINES, S., CASELLE, J. & WARNER, R. (in prep) Patterns beneath the
chaos: dual causes of genetic patchiness in fish recruits. Ecology, in press.
SELKOE, K., VOGEL, A. & AFTAB, B. (2005) Eight polymorphic microsatellite markers
for kelp bass, Paralabrax clathratus, amplified in three multiplex polymerase chain
reaction sets. Molecular Ecology Notes, 5, 127-129.
SELKOE, K., VOGEL, A. & GAINES, S. (in review) Effects of ephemeral circulation on
recruitment and connectivity of nearshore fishes spanning the US-Mexican border.
Marine Ecology Progress Series.
STEPIEN, C., ROSENBLATTT, R. & BARGMEYER, B. (2001) Phylogeography of the
spotted sand bass, Paralabrax maculatorfasciatus: divergence of Gulf of California
and Pacific Coast populations. Evolution, 55, 1852-1862.
SWEARER, S. E., CASELLE, J. E., LEA, D. W. & WARNER, R. R. (1999) Larval
retention and recruitment in an island population of a coral-reef fish. Nature, 402,
799-802.
TAKAGI, M., CHOW, S., OKAMURA, T., SCHOLEY, V., NAKAZAWA, A.,
MARGULES, D., WEXLER, J. & TANIGUCHI, N. (2003) Mendelian inheritance
and variation of four microsatellite DNA markers in the yellowfin tuna Thunnus
albacares. Fisheries Science, 69, 1306-1308.
THOMSON, C. (2001) Human Ecosystem Dimension. IN LEET, W., DEWEES, C.,
KLINGBEIL, R. & LARSON, E. (Eds.) California's living marine resources: A
status report., California Department of Fish and Game.
WAPLES, R. (1998) Seperating the wheat from the chaff: patterns of genetic differentiation
in high gene flow species. Journal of Heredity, 98, 438-450.
WARD, R., WOODWARK, M. & SKIBINSKI, D. (1994) A comparison of genetic diversity
levels in marine, freshwater and anadromous fish. Journal of Fish Biology, 44, 213-
232.
WATTS, R., JOHNSON, M. & BLACK, R. (1990) Effects of recruitment on genetic
patchiness in the urchin Echinometra mathaei in Western Australia. Marine Biology,
105`, 145-151.
9
Chapter 1: The influences of sample size and population differentiation on statistical
power and estimation of F
ST
in marine populations
Abstract
Although migration potentials in the marine environment are typically high, analyses
frequently find genetic structure. Because of difficulties in defining the biological
importance of differentiation and limited use of power analyses, positive results may not be
reflective of biologically-relevant structure. We present a simulation model for calculating
the influence of sample size and population differentiation on statistical power and
estimation of structure. Results quantitatively demonstrate how increased sample size and
differentiation increase power and reduce the range of structure estimates for two marine fish
species: Paralabrax clathratus and Clinocottus analis. Bias against reports of homogeneity
and the need to avoid both under- and over-sampling are discussed in light of these results.
Introduction
Many marine organisms have low levels of genetic structure over large areas because of
pelagic lifecycles and dispersive larvae (Hedgecock, 1994, Ward et al., 1994). Even with
minimal differences, there is a consensus in the marine genetics community that the structure
is significant. We surveyed five journals from 2001 to 2005 that report marine genetic
analyses: “Marine Biology,” “Marine Ecology Progress Series,” “Evolution,” “Canadian
Journal of Fisheries and Aquatic Science” and “Molecular Ecology.” A positive result was
found in 88.5% (23/26) of the articles on species with poor dispersal (because of parental
care, no larval stage, or use of restricted habitats such as estuaries). Interestingly, this result
is no different than the 80% (96/120) of studies on species with dispersive larvae that had
significant structure (binomial distribution with a probability of 0.8; p=0.0841) (Figure 1).
10
Figure 1. Survey of marine population genetics articles in five journals for conclusions of
significant population structure. The percentage of articles on marine species with
planktonic larval stages that find significant structure is shown. The column “Total”
represents percentages based on summed article counts and averaged 80.0% over the five
year period surveyed.
11
Statistical significance though, is not proof of biological importance (Waples, 1998, Hedrick,
1999). Waples (1998) demonstrated how only the discovery of biologically meaningful
structure, as opposed to rejections of the null hypothesis because of Type I error, non-
random sampling or biologically insignificant structure, has merit. There has been no formal
treatment of non-random sampling and the finding of biologically meaningless structure
(Waples, 1998), although more recent work on quantitatively defining populations is highly
relevant (Waples and Gaggiotti, 2006). Of immediate importance is the fact that is difficult
to define an “effect size” (Cohen, 1988) that is universally indicative of genetic structure.
Population differentiation is most often defined by estimates of F
ST
, a parametric value that
ranges from zero to one and is based on levels of inbreeding (Wright, 1951). Estimators or
analogs of F
ST
equate inbreeding and coancestry to allele frequencies (Cockerham, 1969) or
reformulate the basis of fixation indices (Nei, 1977). The maximum estimate of F
ST
though,
is reduced by increasingly high heterozygosities (Hedrick, 2005). The value of F
ST
can also
be influenced by a diversity of other factors beyond migration, including among others,
selection, mutation and effective population size (Neigel, 2002). Further, in the analysis of
high gene flow species such as many marine organisms, error and bias can reduce the
accuracy of structure estimates (Waples, 1998, Neigel, 2002).
Without a definable effect size, a variety of solutions have been proposed for equating
biological to statistical significance, including repetitive sampling (Bentzen, 1998, Waples,
1998) and organism-specific hypotheses (Waples, 1998, Waples and Gaggiotti, 2006).
Reports of geographic structure with significant F
ST
estimates of 0.01 or below
12
(Knutsen et al., 2003, McPherson et al., 2003) continue to be unsettling though. Considering
that populations of finite size will show some structure (Waples, 1998), these low estimates
raise the question of whether there are F
ST
values beyond zero indicative of homogeneity.
Ultimately, two considerations can help studies without well-defined effect sizes. First, if a
biological rationale is part of the hypothesis testing, differentiation values can be judged
more easily (Waples, 1998). One example of research that biologically defines limited
genetic structure is centered on temporal genetic variance (Johnson and Black, 1982). Small
estimates of F
ST
are regularly recorded in this context; (Pujolar et al., 2006) describe average
differences of only 0.0036, (Li and Hedgecock, 1998) find values that average 0.0076 in the
samples of the Pacific oyster (Crassostrea gigas), and (Planes and Lenfant, 2002) note a
global value of 0.0168 between year classes of the white sea bream (Diplodus sargus).
Second, sampling power must be understood (Waples and Gaggiotti, 2006). In general,
differentiation is significant when a test’s p-value is less than !, the likelihood that a result is
a false positive. As sample size becomes larger, significance can be attained with smaller
population differences. This relationship between sample size and significance is balanced
by ", the likelihood that a result is a false negative. If power (1-") is too high because of
excessively large sample sizes, a significant yet biologically irrelevant result can occur.
Although power is a fundamental statistical concept and widely discussed in other fields,
such as behavioral science (Jennions and Møller, 2003, Nakagawa and Foster, 2004),
ecology (Hayes and Steidl, 1997, Thomas, 1997, Jennions and Møller, 2002a, Jennions and
Møller, 2002b, Reed et al., 2002), evolution (Anisimova et al., 2001, Anisimova et al., 2002,
13
Jennions and Møller, 2002a, Jennions and Møller, 2002b, Leebens-Mack and dePamphilis,
2002, Paetkau et al., 2004), genomics (Eddy, 2005), parentage analysis (Neff et al., 2000)
and medicine (Gould, 2001, Grossman and MacKenzie, 2005), it is not systematically
addressed in marine genetics. Sampling might continue until an estimate of diversity can be
made (Cope, 2004), the sampling variance correction is below 0.01 (Waples, 1998), or the
effect size of another species is reached (Schrey and Heist, 2003). Individualized simulation
programs have been created (e.g., Buonaccorsi et al., 2001); it is only recently that more
generalized methods have been promoted (Ryman et al., 2006, Gilbert-Horvath et al., 2006).
The results presented here are an effort to improve awareness of the influences of sample
sizes and population structure on statistical power and estimation of F
ST
. It is an extension
of the discussion on bias by Waples (1998), who encouraged increasing sample sizes and
marker numbers and multi-year sampling to improve analysis of high gene flow species.
Our methodology involves the use of simulations, the details of which are explained in an
effort to promote more standardized evaluation of the statistical limits of sampling.
Materials and Methods
Overview
In marine population genetics, ! is frequently tested through the creation of a null
distribution by genotype permutation with one-tailed significance determined by the
percentage of permutation results that exceed the original test value (e.g. Excoffier et al.,
1992, Belkhir et al., 2004). Dizon et al. (1995) used a similar procedure to calculate power
(1-") by bootstrapping each sample location within itself. Power was considered to be the
percentage of comparisons between bootstrapped samples that were significant.
14
As noted by Dizon et al. (1995), this method does not accurately adjust for sampling
variance. Especially for small samples, allelic richness will be low as compared to the “true
population,” and frequency values will be poorly estimated. Because (1) the sample cannot
be completely representative of the population and (2) no new data have been generated, the
method is actually more an inversion of ! then a calculation of ". Finding this type of
observed power is in fact a pointless calculation, because of its 1:1 relationship with the
already determined, random p-value (Hoenig and Heisey, 2001).
To avoid this logical pitfall, the analysis presented here uses model populations simulated
from empirical data. It is a method similar to the simulations used in Ryman et al. (2006)
(they used purely hypothetical data) and inspired by Cervus© (Marshall et al., 1998), a
program for parentage analysis. That program creates distributions of # values (the
difference between LOD scores of the two most likely candidates) based on either a correct
or incorrect assignment of a simulated parent. Delta scores from actual tests are later given
confidence values based on comparison to the simulated distributions. The strength of the
approach is the use of inherent variance to adjust significance thresholds.
We combined the simulation approach with the Dizon et al. (1995) goal of understanding
Type II error. A simulated population was initiated with genotype scores from empirical
samples but expanded to 100,000 individuals. For markers with neutral distributions, new,
hypothetical alleles were added to the population as determined by $
k
(Ewens, 1972); this
process in effect filled out the tail of the neutral distribution. For markers that did not follow
a neutral distribution, allelic richness was estimated with allele accumulation curves.
15
Differentiation occurred with re-assortment of alleles between two simulated populations.
The process was repeated until populations reached an “F
ST
goal” as measured by %
ST
(Excoffier et al., 1992), G
ST
(Nei and Chesser, 1983) or $ (Weir and Cockerham, 1984).
Because this technique did not use “drift”, it maintained the allelic diversity anticipated in
large populations. The method also avoided estimator bias (e.g., Raufaste and Bonhomme,
2000), which might occur had the method utilized the relationships between effective size,
drift and generation number. The results presented here therefore describe the influence of
sample size and differentiation on F
ST
estimates, but not on F
ST
itself. After populations
were differented to six “F
ST
goals,” samples of seven different sizes were repeatedly
collected. Significance for each sample was calculated through permutation; Power was
estimated to be the percentage of significant p-values.
Programming Tools
All programming was performed in Matlab 7.1 SP3 with the statistics (v.5.2) and curve
fitting (v.1.1.5) toolboxes; .m files and directions for use are available from the first author.
The programming took advantage of MBEToolbox (Cai et al., 2005) (available at
http://www.pmarneffei.hku.hk/mbetoolbox/default.asp), a previously developed tree-analysis
program that can import .fasta files and calculate genetic distances.
Analysis of Sequence Data
Mitochondrial control region sequences from two sources were used: 105 sequences from
Paralabrax clathratus (Genbank accession numbers DQ192295-DQ192399, Vogel and
Edmands, in prep-a), a temperate serranid found along the California and Baja California
16
coasts (Eschmeyer et al., 1983) and 60 sequences from Clinocottus analis (Genbank
accession numbers EF077714-EF077773, Vogel and Edmands, in prep-a), a tidepool sculpin
with roughly the same range as P. clathratus (Eschmeyer et al., 1983).
Each group of sequences was expanded to a hypothetical population of 100,000 alleles.
Because allele distributions were non-neutral according to Tajima’s D (Tajima, 1989) (Table
1), simulated allelic richness was estimated with an exponential curve based on a empirical
data plot of the mean number of alleles (10,000 samples without replacement) (X-axis)
versus sample size (Y-axis). Using that curve, sample sizes that corresponded to the
stepwise increase towards the final allele number were also determined. This method
assumes that historic effective sizes are large and do not limit allelic richness. We calculated
effective population sizes of 208,333 for Paralabrax clathratus and 443,826 for Clinocottus
analis with “Fluctuate” (Kuhner et al., 1998), indicating that the utilization of an expanded
number of alleles is appropriate.
Populations were increased by randomly duplicating one individual at a time. When a
sample size that corresponded to an integer increase allele number was reached, a random
polymorphism with equal probabilities for each of the four nucleic acids and a
transition/transversion ration of 4:1 was introduced. Simulated richness values are presented
in Table 1. To utilize %
ST
(Excoffier et al., 1992) in the simulations, pairwise haplotype
distances were calculated for all alleles. The statistic %
ST
is analogous to F
ST
, but
incorporates haplotype distance information to improve resolution of structure.
17
Table 1. Simulated values of richness for two mitochondrial control region (CR) and seven
microsatellite (MS) markers. Neutrality of sequence data was determined in ARLEQUIN
(Excoffier et al., 2005) with Tajima’s D (Tajima, 1989); a Ewens-Watterson test of
homozygosity (F) (Hartl and Clark, 1989) was performed in “Enumerate” (Slatkin 1996)
with microsatellite data. Allelic richness is given for empirical data (pre-adjustment) and
after simulation with an allele accumulation curve or Ewens $
k
(Ewens, 1972). Except for
microsatellite Mbo66 ($
k
was set to 1), $
k
was rounded to the nearest integer in the creation
of simulated populations.
18
The simulated population was differentiated from a copy of itself until it reached F
ST
goals
as determined by %
ST
(Appendix 1). The process involved defining a maximum “switch
number;” the simulation exchanged randomly-chosen alleles between the populations in a
quantity determined by a random integer value between one and that maximum. Switching
was repeated in both directions to maintain equal-sized populations.
For each F
ST
goal, 25 population pairs were created. Five, 15, 20, 30, 50 and 100 haplotypes
per population were sampled from each pair (Appendix 1). To describe the influence of
sample size on determining “true” population differentiation, %
ST
values were calculated for
each sample set. Significance for a set was determined by 1000 permutations (as in
Excoffier et al., 1992). The sampling of each population pair was repeated 1000 times, for a
total of 25,000 samples collected for each size. The estimate of power was the percentage of
significant results.
A confidence interval was created with batch samples (e.g., Excoffier et al., 2005, see user
manual). The 25,000 samples were divided into 250, 100-sample batches, for each of which
power was recalculated. Because p-values range between zero and one, their values follow a
"–distribution. A function in the statistics toolbox of Matlab ® (betafit.m) was utilized to
perform a maximum-likelihood estimation of the ! and " parameters. The 95% confidence
interval of the ! and " parameters was calculated with a &
2
distribution according to the
equations:
!
(n"1)*[#|$]
1"#/2
2
%
[df =n"1]
,
(n"1)*[#|$]
#/2
2
%
[df =n"1]
&
'
(
(
)
*
+
+
(an adaptation of equation 16 in
(Waples, 1989)). The lower value of the power estimate’s confidence interval was taken as
the 2.5 percentile of a hypothetical "–distribution created from the lower estimate of the !
19
parameter and the upper estimate of the " parameter. The upper value of the power
estimate’s confidence interval was the inverse: the 97.5 percentile for the upper estimate of
the ! parameter and the lower estimate of the " parameter.
A Spearman rank correlation was utilized to test confidence interval size against a power
estimate’s minimum distance from zero or one. After an inverse sine transformation of the
power estimates, the influence of sample size and differentiation was ascertained with an
Analysis of Variance (ANOVA). Wilcoxon signed rank tests were used to test if power
estimates were significantly different between Paralabrax clathratus and Clinocottus analis
data, and between data calculated with G
ST
and $. An ANOVA was also employed to test
the influence of sample size and differentiation on the range, skewness and kurtosis
(measure of outlier values) of F
ST
estimates. Significance of skewness was determined with
a T-Test for mean ' 0; the same test was implemented to establish whether mean kurtosis
was >3. Lastly, estimate bias was analyzed by calculating the mean and median sample size
needed for estimates to be a value, on average, that was at least 90% of the “true” level of
differentiation. Direction of bias was tested with a Cox-Stuart non-parametric test for trend
(Cox and Stuart, 1955) of the mean and median values of the estimates.
Analysis of Microsatellite Data
Analysis was performed on data collected from 2862 individuals of Paralabrax clathratus
(Vogel and Edmands, in prep-b) with seven microsatellite loci (published in Selkoe et al.,
2005). The procedure was identical to that done with the above mitochondrial sequence
data, but used different genetic structure statistics and included manipulation of multiple, co-
dominant loci. Because the microsatellites had neutral distributions according to a Ewens-
20
Watterson test of homozygosity (F) (Hartl and Clark, 1989) calculated with “Enumerate”
(Slatkin, 1996), we utilized Ewens’ $
k
(Ewens, 1972), calculated in ARLEQUIN 3.01
(Excoffier et al., 2005), to estimate allele numbers (Table 1). Using integer values of $
k
(rounded to improve calculation time), the expected number of alleles from 100,000
individuals was calculated by:
!
E(k)="k
1
"k +1
0
n-1
#
(Ewens, 1972). Microsatellite locus
Mbo66, with a calculated $
k
of 0.23, was simulated with a $
k
value of 1.
Twenty-five population pairs were differentiated by G
ST
(Nei and Chesser, 1983) or $ (Weir
and Cockerham, 1984). Random union of gametes were used in differentiation (G
ST
: Nei
and Chesser, 1983, equations 15 and 16; $: Reynolds et al., 1983). During testing of
sampling, H
O
was incorporated (G
ST
: Nei and Chesser, 1983, equations 5, 9 and 11; $: Weir
and Cockerham, 1984, equations 2, 3 and 4, $
W
). Assuming Hardy-Weinberg equilibrium,
alleles were individually sampled but built into genotypes of seven, diploid loci.
Results
Sequence results: Analysis of power estimations for !
ST
Complete results are presented in Table 2. Power estimates for Paralabrax clathratus data
ranged from 0.035 to 1 and for Clinocottus analis data, 0.002 to 1. Although true power is
between 0.05 and 1 with an ! = 0.05 (Gerard et al., 1998), the simulation process has
random inaccuracies and eleven estimates were lower than 0.05 (confidence intervals
included 0.05). To double-check functionality and provide a landmark (Cockerham and
Weir, 1993), we ran 25 simulations for populations of P. clathratus with %
ST
values = 0.000.
Power estimates were between 0.034 and 0.051 and %
ST
estimates ranged from –0.0023 to 0.
21
Table 2. Power results from sampling of the mitochondrial control region for differentiated
simulated populations. For each drift goal, 25 population pairs were created, each of which
was sampled 1000 times for each of the sample sizes (for a total of 25,000 samples).
22
Table 2. Continued.
23
The power estimations reveal a number of interesting patterns. First, middle-range power
estimates were the most ambiguous; correlations of confidence interval size and the distance
from 0 or 1 were found in both Paralabrax clathratus ((=0.9355, p<0.0001) and Clinocottus
analis data ((=0.9084, p<0.0001). Second, significant influence of sample size and
population differentiation on the point estimate of power was found for both species (Table
3). Third, power was affected by allelic richness and significantly higher for data from
Paralabrax clathratus (p<0.0001, Figures 2-3).
Sequence results: Analysis of estimation of !
ST
All data are in Table 4. Figure 4 is a plot of sample size versus estimated %
ST
for simulated
Paralabrax clathratus populations drifted to %
ST
= 0.05. Three patterns can be seen in the
plot. First, as seen in the 95% range, there was a significant, negative trend in the estimate
ranges as sample size increased (Table 5). In the Figure 4 example, a sample size of 5 had a
95% range of 0.3042, whereas a sample size of 100 had a reduced range of 0.0540. Increase
in population differentiation also decreased the estimate ranges (Table 5).
Second, median %
ST
estimates were consistently below the mean, caused by right-skewed
distributions in Paralabrax clathratus data (mean skewness=0.787, t=15.58, p<0.0001) and
Clinocottus analis data (mean skewness=0.901, t=13.15, p<0.0001). Skewness decreased as
sample size or population differentiation increased (Table 5). There was also a significant
number of outlier values of high %
ST
estimates in P. clathratus data (mean kurtosis=4.20,
t=9.18, p<0.0001) and C. analis data (mean kurtosis=4.49, t=7.61, p<0.0001), which was
reduced by increased sample size and population differentiation (Table 5).
24
Table 3. Results from ANOVA tests of the influence of sample size and population
differentiation on the point estimate of power. Data were normalized with an inverse sine
transformation.
Sample Size Population Differentiation
F d.f p-value F d.f p-value
P. clathratus 21.04 6 <0.0001 51.96 5 <0.0001
C. analis 19.38 6 <0.0001 32.29 5 <0.0001
G
ST
10.26 6 <0.0001 30.43 5 <0.0001
$ 7.80 6 <0.0001 18.74 5 <0.0001
25
Table 4. Results describing the estimation of %
ST
from sampling of differentiated simulated
populations. For each drift goal, 25 population pairs were created, each of which was
sampled 1000 times for each sample size (25,000 samples). The 95% Limits of Data
represent the lower and upper %
ST
estimates that bracket 95% of the data (23,750 samples).
26
Table 4. Continued.
27
Table 5. Results from 2-way ANOVA analyses on the effects of sample size and population
differentiation on estimation of F
ST
. Tests included effects on the range, skewness and
kurtosis of F
ST
estimates. Data were normalized with an inverse sine transformation; for %
ST
,
the two species were treated as replicates.
Sample Size Population Differentiation
Test F d.f. p-value F d.f. p-value
%
ST
range 66.48 6 <0.0001 9.16 5 <0.0001
skewness 4.19 6 0.0022 64.20 5 <0.0001
kurtosis 5.22 6 0.0004 59.94 5 <0.0001
G
ST
range 183.66 6 <0.0001 12.00 5 <0.0001
skewness 3.48 6 0.0100 13.68 5 <0.0001
kurtosis 1.88 6 0.1177 14.85 5 <0.0001
$ range 235.96 6 <0.0001 4.83 5 0.0023
skewness 4.19 6 0.0060 8.97 5 <0.0001
kurtosis 4.00 6 0.0040 6.68 5 <0.0001
28
Figure 2. Plot comparing power sampling of two hypothetical populations which were
drifted to a %
ST
goal of 0.01. Power estimates represent 25,000 samples taken for each
sample size from Paralabrax clathratus and Clinocottus analis simulated data.
29
Figure 3. Plot comparing power sampling of two hypothetical populations which were
drifted to a %
ST
goal of 0.10. Power estimates represent 25,000 samples taken for each
sample size from Paralabrax clathratus and Clinocottus analis simulated data.
30
Figure 4. Plot of sample size versus estimated %
ST
for 25,000 samples (each sample size)
from Paralabrax clathratus mitochondrial sequence data. Plot includes mean, median and
95% range of %
ST
estimates for simulated populations that were differentiated to %
ST
= 0.05.
31
The third pattern was mean and median %
ST
estimates asymptotically approaching the
differentiation value from below. In Table 6 are sample sizes needed to achieve a mean or
median of the estimate of 90% of the population value. Despite small numbers of data
points in the Cox-Stuart tests (number of samples sizes (7) minus 1; significance could only
be achieved with stepwise increase in all values), slopes were significantly positive for mean
estimates in 3/6 levels of differentiation for Paralabrax clathratus data and all levels in
Clinocottus analis data. For median values, there were significantly positive slopes for 5/6
differentiation values in P. clathratus data and 4/6 in C. analis data.
Microsatellite results: Analysis of power estimations for G
ST
and "
Results were nearly identical to those found with sequence data; complete results are in
Table 7. Power estimates ranged from 0.053 to 1 with G
ST
data and 0.051 to 1 for data
analyzed with $. Like the mitochondrial results, confidence intervals were correlated with
distance from the extremes for G
ST
data ((=0.9432, p<0.0001) and $ data ((=0.8964,
p<0.0001). Sample size and population differentiation significantly influenced point
estimates of power for both G
ST
and $ (Table 3).
Data analyzed with $ had methodically lower power estimates (p<0.0001) that matched
estimates from G
ST
populations that were half as differentiated (Figure 5). The power
estimate for 5 samples taken from $ = 0.02 populations for example, was 0.189. This
matched the power estimate (0.187) for 5 samples taken from populations that were
differentiated to G
ST
= 0.01.
32
Table 6. Sample sizes needed to achieve mean or median estimates of (1) %
ST
that were at
least 90% of the differentiation value for the two hypothetical drift populations of
Paralabrax clathratus and Clinocottus analis (2) GST and $ that were at least 90% of the
differentiation value. Sample sizes of >100 indicate that the 90% value was not reached for
the sample sizes tested in this study.
33
Table 7. Power results from sampling of microsatellite markers for differentiated simulated
populations. For each drift goal, 25 population pairs were created, each of which was
sampled 1000 times for each of the sample sizes (for a total of 25,000 samples). Two
genetic measures were used: G
ST
and $.
34
Table 7. Continued.
35
Figure 5. Plot comparing sample size versus resolution of estimated power for data from
Paralabrax clathratus measured with G
ST
and $. Power estimates from samples populations
differentiated to 0.0005, 0.001, and 0.01 represent G
ST
; a set of staggered values of 0.001,
0.002 and 0.02 represent $. The staggered values of $ match the lower values of G
ST
. The
95% confidence intervals have been removed to reduce clutter in the plot.
36
Microsatellite results: Analysis of estimation of G
ST
and "
Results can be found in Table 8. Estimate ranges were significantly influenced by both
increased sample size and population differentiation (Table 5). There was significant right-
handed skewness to the data for G
ST
(mean skewness=0.703, t=38.06, p<0.0001) and $
(mean skewness=0.709, t=51.43, p<0.0001) which was reduced by increased sample size
and population differentiation (Table 5). Kurtosis was significant for G
ST
data (mean
kurtosis=4.01, t=21.26, p<0.0001) and $ data (mean kurtosis=4.00, t=26.45, p<0.0001).
Kurtosis for G
ST
data was related to population differentiation but not sample size; kurtosis
was related to both for $ data (Table 5).
An area where these data were different than %
ST
estimates regarded the sample size needed
for estimation of population differentiation within 90% of the true level (Table 6). Mean
estimates matched differentiation values with sample sizes of either 5 or 10 diploid
individuals for both G
ST
and $ data. Median estimates required larger sample sizes, but
could anticipate structure with these small sample sizes if differentiation was greater (e.g.,
G
ST
and $ = 0.02). None of the Cox Stuart tests were significant for mean estimates but all
were significant for median values, a pattern attributable to change in skew and kurtosis.
Discussion
Sampling power as analyzed by this model
An immediate trend in the results is the preponderance of large confidence intervals for
power (Tables 2 and 7). As power has been estimated, confidence intervals are an important
consideration and an area in which we have expanded upon the work of Ryman et al. (2006)
and (Waples and Gaggiotti, 2006).
37
Table 8. Results describing the estimation of G
ST
and $ with sampling of differentiated
simulated populations. For each drift goal, 25 population pairs were created, each of which
was sampled 1000 times for each of the sample sizes (25,000 samples). The 95% Limits of
Data represent the lower and upper estimates that bracket 95% of the data (23,750 samples).
38
Table 8. Continued.
39
A practical consequence of this pattern is that error is highest for the power estimates in
which one would be most interested, 0.5 to 0.8 (Gerard et al., 1998). Nonetheless, sampling
strategy would still benefit from knowing a sample size had low power. Small sample sizes
can be useful for finding evidence of strong barriers to dispersal, but care would have to be
taken not to over-dramatize more subtle differences. Understanding a sampling-scheme’s
power could also be used to develop standardized sample sizes or define the number of
markers to develop from a microsatellite clone library.
There are a number of other points to be taken from these data. First, large sample sizes are
needed to explore the lowest levels of genetic structure, as noted by Waples, (1998) and
Ryman et al., (2006), and such as found in Feldheim et al. (2001). With Paralabrax
clathratus for example, analysis of structure with %
ST
! 0.02 would require sample sizes of
100 to have an estimated power of at least 0.8. A more modest sample size of ~25 could be
used to test for structure of %
ST
" 0.05.
Second, power estimates for $ were half as strong as those for G
ST
. Other simulations
(Raufaste and Bonhomme, 2000) have indicated that $ as calculated by Weir and
Cockerham has higher variance in its estimations than $ as calculated by (Robertson and
Hill, 1984). In general, the range of $ estimates was about twice that of G
ST
(Table 8).
Because $, for any single estimation, has a reduced probability of correctly estimating
differentiation, its power was also reduced by a factor of two.
40
Third, modeling strong differences with sequence data that empirically showed little
difference may not be appropriate. Unsampled populations for example, might have highly
divergent allele complements, something we did not test. Because of inclusion of more
divergent or unique alleles, power would by higher. In that case, power could be improved
or alternatively, become too high if analysis was of biologically irrelevant structure.
Estimation of F
ST
The 95% limits of the F
ST
estimates with small sample sizes are wide, and for low levels of
differentiation encompass values an order of magnitude different than the “true” F
ST
. Single
sample F
ST
estimations have a low probability of documenting the correct F
ST
(Chakraborty
and Danker-Hopfe, 1991). Our results on right-handed skewness and kurtosis present an
intriguing aspect relevant to this point. Because F
ST
estimates become more normally
distributed as differentiation and sample size increase, the likelihood of a large/spurious
result is higher for small sample sizes and low levels of differentiation. In studies that find
low structure with significance driven by a small subset of markers or locations, drift and
selection continue to be reasonable possibilities, but sampling variance cannot be ignored.
We have also found evidence that %
ST
was influenced by sample size (Table 6, Figure 4) but
G
ST
and $ are non-biased estimators (Table 6) (e.g., Nei and Chesser, 1983, Weir and
Cockerham, 1984). Despite disagreement on the logical basis of the estimators (Weir and
Cockerham, 1984, Chakraborty and Danker-Hopfe, 1991), both give almost exactly the same
estimates (Cockerham and Weir, 1986). We performed 25 samplings of hypothetical data
(differentiated to G
ST
/$=0.001) and found that ratios of the estimations had a mean of
1.0004 (t-test for mean different than 1; t=0.37495, p=0.7110).
41
All of these results support the use of either confidence intervals for F
ST
or simulation of
95% limits, as performed here. Confidence intervals require replicates, available with
analysis of multiple markers. For a single marker, the only option might be temporal
sampling; with large sample sizes, a batch size approach might also work. Nurbaev and
Balanovskaia, (1998) found that F
ST
estimates follow a "–distribution, which presents an
alternative in confidence interval creation. Contingency and exact tests can also be
implemented, with less bias than confidence intervals (Waples and Gaggiotti, 2006).
Influence of allelic richness
Simulation methods frequently assume effective population sizes of less than 1000 and
appropriately low numbers of alleles (e.g., Waples and Gaggiotti, 2006, Ryman et al., 2006).
Although allelic richness in mitochondrial markers does not follow population size (Bazin et
al., 2006) and N
e
/N may be small (Hoarau et al., 2005), it is empirically evident that historic
effective sizes in many marine species are large enough that allelic richness is high (e.g.,
Rocha-Olivares and Vetter, 1999, Cope, 2004). In the case of the Paralabrax clathratus
data used here, there were already 53 alleles in 105 sequences. Modeling differentiation for
these species is therefore improved with an increased number of alleles.
We have not explicitly tested the effects of allelic richness on estimation of power and F
ST
,
but it has been noted that markers with larger numbers of alleles are more powerful
estimators of structure (Goudet et al., 1996, Hedrick, 1999). Although it is doubtful that
poor sampling of rare alleles could significantly change estimates of power, a large
difference in anticipated and true allelic richness could be important. For example, the
42
Paralabrax clathratus data, which was simulated with an allelic richness of 1414 alleles,
consistently had higher power than the Clinocottus analis data with only 373 alleles. A
series of simulations (data not shown) found that use of the confidence interval of Ewen’s $
k
did not affect the confidence intervals for power, and that power was only significantly
influenced when allelic richness was changed by roughly an order of magnitude. This was
also found to be true for the 95% limits of the F
ST
estimates.
Research considerations of power and F
ST
estimation
One of the most contentious issues with statistical power is that of retrospective analysis
(Gerard et al., 1998, Hoenig and Heisey, 2001). The need to flesh out a null hypothesis
result is well understood, but resampling is only an inversion of the p value. Ideally, a
power analysis should analyze a pilot study to improve sampling in the main collection. At a
minimum, the method detailed here permits exploration of the limits of an analysis.
The problem with “exploring data” is that it should not be driven by a p-value. Instead of an
effort to define the smallest, statistically significant effect size, the objective of a power
analysis should be driven by the goal of identifying biologically-relevant landmarks to the
research question (Gerard et al. [1998] make a similar argument). Waples and Gaggiotti,
(2006) present a conceptual framework for using migration rates to identify biologically
relevant gene flow, but point out that those values must still be determined on a case by case
basis. If it is hard to universally define appropriate %
ST
, G
ST
and $ values to biological
events, then it is also difficult to present generalized suggestions. We have a number of
ideas, which in large part reinforce previous suggestions (Waples, 1998, Ryman et al., 2006,
Waples and Gaggiotti, 2006).
43
First, a project designed solely to “find structure” is preliminary in nature and might be
better built around specific foci. These include suspect phylogeographic barriers, areas
where clines and isolation by distance could be important, or questions about temporal
variance. Were an isolation by distance pattern the main question for example, sampling
could follow Palumbi's (2003) suggestion of comparing paired populations with increasingly
large distances (as done by Fauvelot and Planes, 2002). If tangential analyses were
performed or an unexpected or unusual pattern found, it would not be the entire focus of the
publication. Rather, it would be a discussion point in “future directions.”
More consistently designed projects would make null results more palatable. Finding no
evidence of isolation by distance in the face of an active hypothesis test is a more logically
and statistically sound result than testing for patterns with no a priori hypothesis during the
sampling. It also presents the opportunity to better gauge the broader question of whether
there is structure in the ocean.
Second, sample sizes of the articles in our analysis of five journals indicate that typical
studies have high statistical power. Average sample size was 36.6 individuals per location;
84 articles had average sizes greater than 25, 27 greater than 50, and 8 greater than 100 (six
articles were not counted because of difficulty in determining sample sizes). These are all
sizes, assuming that our data can be extrapolated, within the range needed to find significant
structure of values that are regularly found in the marine realm: %
ST
< 0.1 or 0.002 ) G
ST
*
0.2. If power is high and statistical significance achievable for almost all F
ST
values, then it
is immediately important for effect sizes to be biologically definable. Working to define
important limits of structure a priori (e.g., Waples and Gaggiotti, 2006) is a better approach
44
than a post hoc defense of “whatever was found.” Without biological criteria, definition of
biological importance through p-values will be heavily influenced by easy-to-achieve sample
sizes and marker numbers. Power analysis, such as the approach presented here and those
described by Gilbert-Horvath et al. (2006) and Ryman et al. (2006), can help organize
sampling regimes to specifically test benchmark effect sizes.
Third, repeatability or evidence of temporal pattern must become more important, especially
if differences are low and have been reported for the first time (noted by (Waples, 1998).
Sampling over multiple years is also the single best method for adjusting for statistical
power that is too strong. From our survey of 149 articles, 29/149 (19.5%) had multi-year
sampling schemes that surveyed temporal variance or stability in geographic patterns.
Without ample caveats, it is dangerous to move into discussions about a pattern (such as the
influences of glaciation or management for example) if there is no evidence of temporal
consistency. Interpretation can change with repeated sampling; only after five studies on the
New Zealand greenshell mussel (Perna canaliculus) could a north/south barrier became the
most parsimonious explanation (Star et al., 2003).
Acknowledgements
This research was supported by Teaching Assistantships from the University of Southern
California and Rose Hills Summer Fellowships from the Wrigley Institute for Environmental
Studies. We thank M. Neumann and B. Jones for their advice with coding in Matlab®.
45
Chapter 1 References
ANISIMOVA, M., BIELAWSKI, J. & YANG, Z. (2001) Accuracy and power of the
likelihood ratio test in detecting adaptive molecular evolution. Molecular Biology
and Evolution, 18, 1585-1592.
ANISIMOVA, M., BIELAWSKI, J. & YANG, Z. (2002) Accuracy and power of Bayes
prediction of amino acid sites under positive selection. Molecular Biology and
Evolution, 19, 950-958.
BAZIN, E., GLÉMIN, S. & GALTIER, N. (2006) Population size does not influence
mitochondrial genetic diversity in animals. Science, 312, 570-572.
BELKHIR, K., BORSA, P., CHIKHI, L., RAUFASTE, N. & BONHOMME, F. (2004)
Genetix 4.05, logiciel sous Windows TM pour la génétique des populations. IN
LABORATOIRE GÉNOME, P., INTERACTIONS, CNRS UMR 5171 (Ed.)
Montpellier (France), Université de Montpellier II.
BENTZEN, P. (1998) Seeking evidence of local stock structure using molecular genetic
methods. IN VON HERBING, I., KORNFIELD, I., TUPPER, M. & WILSON, J.
(Eds.) The implications of localized fishery stocks. Portland, Maine, Natural
resource, agriculture, and engineering service.
BUONACCORSI, V., STARKEY, E. & GRAVES, J. (2001) Mitochondrial and nuclear
DNA analysis of population subdivision among young-of-the-year Spanish mackerel
(Scomberomorus maculatus) from the western Atlantic and Gulf of Mexico. Marine
Biology, 138, 37-45.
CAI, J., SMITH, D., XIA, X. & YUEN, K. (2005) MBEToobox: a Matlab toolbox for
sequence data analysis in molecular biology and evolution. BMC Bioinformatics, 6,
doi:10.1186/1471-2105-6-64.
CHAKRABORTY, R. & DANKER-HOPFE, H. (1991) Analysis of population structure: a
comparative analysis of different estimators of Wright's fixation indices. IN RAO,
C. & CHAKRABORTY, R. (Eds.) Handbook of statistics 8: Statistical methods in
biological and medical sciences. Amsterdam, the Netherlands, Elsevier Science
Publishers B.V.
COCKERHAM, C. (1969) Variance of gene frequencies. Evolution, 23, 72-84.
COCKERHAM, C. & WEIR, B. (1986) Estimation of inbreeding parameters in stratified
populations. Annals of Human Genetics, 50, 271-281.
COCKERHAM, C. & WEIR, B. (1993) Estimation of gene flow from F-statistics. Evolution,
47, 855-863.
COHEN, J. (1988) Statistical power analysis for the behavioral sciences, Mahwah, New
Jersey, Lawrence Erlbaum Associates.
COPE, J. (2004) Population genetics and phylogeography of the blue rockfish (Sebastes
mystinus) from Washington to California. Canadian Journal of Fisheries and
Aquatic Science, 61, 332-342.
COX, D. & STUART, A. (1955) Some quick tests for trend and dispersion. Biometrika, 42,
80-95.
DIZON, A., TAYLOR, B. & O'CORRY-CROWE, G. (1995) Why statistical power is
necessary to link analyses of molecular variation to decisions about population
structure. American Fisheries Society Symposium, 17, 288-294.
EDDY, S. (2005) A model of the statistical power of comparative genome sequence
analysis. Public Library of Science, Biology, 3, 95-102.
46
ESCHMEYER, W. N., HERALD, E. S. & HAMMANN, H. (1983) A Field Guide to Pacific
Coast Fishes, New York, New York, Houghton Mifflin Company.
EWENS, W. (1972) The sampling theory of selectively neutral alleles. Theoretical
Population Biology, 3, 87-112.
EXCOFFIER, L., LAVAL, L. & SCHNEIDER, S. (2005) Arlequin ver. 3.0: An integrated
software package for population genetics data analysis. Evolutionary Bioinformatics
Online, 1, 47-50.
EXCOFFIER, L., SMOUSE, P. & QUATTRO, J. (1992) Analysis of molecular variance
inferred from metric distances among DNA haplotypes: application to human
mitochondrial DNA restriction data. Genetics, 131, 479-491.
FAUVELOT, C. & PLANES, S. (2002) Understanding origins of present-day genetic
structure in marine fish: biologically or historically driven patterns? Marine Biology,
141, 773-788.
FELDHEIM, K., GRUBER, S. & ASHLEY, M. (2001) Population genetic structure of the
lemon shark (Negaprion brevirostris) in the western Atlantic: DNA microsatellite
variation. Molecular Ecology, 10, 295-303.
GERARD, P., SMITH, D. & WEERAKKODY, G. (1998) Limits of retrospective power
analysis. Journal of Wildlife Management, 62, 801-807.
GILBERT-HORVATH, E., LARSON, R. & GARZA, J. (2006) Temporal recruitment
patterns and gene flow in kelp rockfish (Sebastes atrovirens). Molecular Ecology,
doi: 10.1111/j.1365-294X.2006.03033.x.
GOUDET, J., RAYMOND, M., DE MEEÜS, T. & ROUSSET, F. (1996) Testing
differentiation in diploid populations. Genetics, 144, 1933-1940.
GOULD, A. (2001) Sample size re-estimation: recent developments and practical
considerations. Statistics in Medicine, 20, 2625-2643.
GROSSMAN, J. & MACKENZIE, F. (2005) The randomized controlled trial. Perspectives
in Biology and Medicine, 48, 516-534.
HARTL, D. & CLARK, A. (1989) Principals of Population Genetics, Sunderland, Sinauer
Associates, Inc.
HAYES, J. & STEIDL, R. (1997) Statistical power analysis and amphibian population
trends. Conservation Biology, 11, 273-275.
HEDGECOCK, D. (1994) Temporal and spatial genetic structure of marine animal
populations in the California current. California Cooperative Fisheries
Investigations, Reports, 35, 73-81.
HEDRICK, P. (1999) Perspective: Highly variable loci and their interpretation in evolution
and conservation. Evolution, 53, 313-318.
HEDRICK, P. (2005) A standardized genetic differentiation measure. Evolution, 59, 1633-
1638.
HOARAU, G., BOON, E., JONGMA, D., FERBER, S., PALSSON, J., VAN DER VEER,
H., RIJNSDORP, A., STAM, W. & OLSEN, J. (2005) Low effective population size
and evidence for inbreeding in an overexploited flatfish, plaice, (Pleuronectes
platessa L.). Proceedings of the Royal Society of London B: Biological Sciences,
272, 497-503.
HOENIG, J. & HEISEY, D. (2001) The abuse of power: the pervasive fallacy of power
calculations for data analysis. The American Statistician, 55, 1-6.
JENNIONS, M. & MØLLER, A. (2002a) Publication bias in ecology and evolution: an
empirical assessment using the "trim and fill" method. Biological Reviews, 77, 211-
222.
47
JENNIONS, M. & MØLLER, A. (2002b) Relationships fade with time: a meta-analysis of
temporal trends in publication in ecology and evolution. Proceedings of the Royal
Society of London B: Biological Sciences, 269, 43-48.
JENNIONS, M. & MØLLER, A. (2003) A survey of the statistical power of research in
behavioral ecology and animal behavior. Behavioral Ecology, 14, 438-445.
JOHNSON, M. S. & BLACK, R. (1982) Chaotic genetic patchiness in an intertidal limpet,
Siphonia sp. Marine Biology, 70, 157-164.
KNUTSEN, H., JORDE, P., ANDRE, C. & STENSETH, N. (2003) Fine-scaled
geographical population structuring in a highly mobile marine species: the Atlantic
cod. Molecular Ecology, 12, 385-394.
KUHNER, M., YAMATO, J. & FELSENSTEIN, J. (1998) Maximum likelihood estimation
of population growth rates based on the coalescent. Genetics, 149, 429-434.
LEEBENS-MACK, J. & DEPAMPHILIS, C. (2002) Power analysis of tests for loss of
selective constraint in cave crayfish and nonphotosynthetic plant lineages.
Molecular Biology and Evolution, 19, 1292-1302.
LI, G. & HEDGECOCK, D. (1998) Genetic heterogeneity, detected by PCR-SSCP, among
samples of larval Pacific oysters (Crassostrea gigas) supports the hypothesis of
large variance in reproductive success. Canadian Journal of Fisheries and Aquatic
Science, 55, 1025-1033.
MARSHALL, T., SLATE, J., LEB, K. & PEMBERTON, J. (1998) Statistical confidence for
likelihood-based paternity inference in natural populations. Molecular Ecology, 7,
639-655.
MCPHERSON, A., STEPHENSON, R. & TAGGART, C. (2003) Genetically different
Atlantic herring Clupea harengus spawning waves. Marine Ecology Progress
Series, 247, 303-309.
NAKAGAWA, S. & FOSTER, T. (2004) The case against retrospective statistical power
analyses with an introduction to power analysis. Acta Ethologica, 7, 103-108.
NEFF, B., REPKA, J. & GROSS, M. (2000) Statistical confidence in parentage analysis
with incomplete sampling: how many loci and offspring are needed? Molecular
Ecology, 9, 529-539.
NEI, M. (1977) F-statistics and analysis of gene diversity in subdivided populations. Annals
of Human Genetics, 11, 307-325.
NEI, M. & CHESSER, R. (1983) Estimation of fixation indices and gene diversities. Annals
of Human Genetics, 47, 253-259.
NEIGEL, J. (2002) Is F
ST
obsolete? Conservation Genetics, 3, 167-173.
NURBAEV, S. & BALANOVSKAIA, E. (1998) Interpopulation diversity of the gene pool:
beta distribution of Wright's F
ST
statistics. Russian Journal of Human Genetics, 34,
1004-1008.
PAETKAU, D., SLADE, R., BURDEN, M. & ESTOUP, A. (2004) Genetic assignment
methods for the direct, real-time estimation of migration rate: a simulation-based
exploration of accuracy and power. Molecular Ecology, 13, 55-65.
PALUMBI, S. (2003) Population genetics, demographic connectivity, and the design of
marine reserves. Ecological Applications, 13, S146-S158.
PLANES, S. & LENFANT, P. (2002) Temporal change in the genetic structure between and
within cohorts of a marine fish, Diplodus sargus, induced by a large variance in
individual reproductive success. Molecular Ecology, 11, 1515-1524.
PUJOLAR, J., MAES, G. & VOLCKAERT, F. (2006) Genetic patchiness among recruits in
the European eel Anguilla anguilla. Marine Ecology Progress Series, 307, 209-217.
48
RAUFASTE, N. & BONHOMME, F. (2000) Properties of bias and variance of two
multiallelic estimators of F
ST
. Theoretical Population Biology, 57, 285-296.
REED, J., MILLS, L., DUNNING JR., J., MENGES, E., MCKELVEY, K., FRYE, R.,
BEISSINGER, S., ANSTETT, M. & MILLER, P. (2002) Emerging issues in
population viability analysis. Conservation Biology, 16, 7-19.
REYNOLDS, J., WEIR, B. & COCKERHAM, C. (1983) Estimation of the coancestry
coefficient: basis for a short-term genetic distance. Genetics, 105, 767-779.
ROBERTSON, A. & HILL, W. (1984) Deviations from Hardy-Weinberg proportions:
sampling variances and use in estimation of inbreeding coefficients. Genetics, 107,
703-718.
ROCHA-OLIVARES, A. & VETTER, R. D. (1999) Effects of oceanographic circulation on
the gene flow, genetic structure, and phyogeography of the rosethorn rockfish
(Sebastes helvomaculatus). Canadian Journal of Fisheries and Aquatic Science, 56,
803-813.
RYMAN, N., PALM, S., ANDRÉ, C., CARVALHO, G., DAHLGREN, T., JORDE, P.,
LAIKRE, L., LARSSON, L., PALMÉ, A. & RUZZANTE, D. (2006) Power for
detecting genetic divergence: differences between statistical methods and marker
loci. Molecular Ecology, 15, 2031-2045.
SCHREY, A. & HEIST, E. (2003) Microsatellite analysis of population structure in the
shortfin mako (Isurus oxyrinchus). Canadian Journal of Fisheries and Aquatic
Science, 60, 670-675.
SELKOE, K., VOGEL, A. & AFTAB, B. (2005) Eight polymorphic microsatellite markers
for kelp bass, Paralabrax clathratus, amplified in three multiplex polymerase chain
reaction sets. Molecular Ecology Notes, 5, 127-129.
SLATKIN, M. (1996) A correction to the exact test based on the Ewens sampling
distribution. Genetical Research, 68, 259-260.
STAR, B., APTE, S. & GARDNER, J. (2003) Genetic structuring among populations of the
greenshell mussel Perna canaliculus revealed by analysis of randomly amplified
polymorphic DNA. Marine Ecology Progress Series, 249, 171-182.
TAJIMA, F. (1989) Statistical method for testing the neutral mutation hypothesis by DNA
polymorphism. Genetics, 123, 585-595.
THOMAS, L. (1997) Retrospective power analysis. Conservation Biology, 11, 276-280.
VOGEL, A. & EDMANDS, S. (in prep-a) Population genetic structure in Paralabrax
clathratus and Clinocottus analis across potential barriers to gene flow.
VOGEL, A. & EDMANDS, S. (in prep-b) Temporal and spatial genetic variance in the kelp
bass, Paralabrax clathratus.
WAPLES, R. (1989) A generalized approach for estimating effective population size from
temporal changes in allele frequency. Genetics, 121, 379-391.
WAPLES, R. (1998) Seperating the wheat from the chaff: patterns of genetic differentiation
in high gene flow species. Journal of Heredity, 98, 438-450.
WAPLES, R. & GAGGIOTTI, O. (2006) What is a population? An empirical evaluation of
some genetic methods for identifying the number of gene pools and their degree of
connectivity. Molecular Ecology, 15, 1419-1439.
WARD, R., WOODWARK, M. & SKIBINSKI, D. (1994) A comparison of genetic diversity
levels in marine, freshwater and anadromous fish. Journal of Fish Biology, 44, 213-
232.
WEIR, B. & COCKERHAM, C. (1984) Estimating F-statistics for the analysis of population
structure. Evolution, 38, 1358-1370.
49
WRIGHT, S. (1951) The genetical structure of populations. Annals of Eugenics, 15, 323-
353.
50
Chapter 2: Population genetic structure in Paralabrax clathratus and Clinocottus analis
across potential barriers to gene flow
Abstract
Barriers in marine systems limit migration and are important to fisheries management, but
many do not restrict gene flow universally. The kelp bass, Paralabrax clathratus, and the
woolly sculpin, Clinocottus analis, were surveyed through their ranges and across Punta
Eugenia, a putative break on the Pacific side of the Baja Peninsula. Average F
ST
values were
low and non-significant (0.0144 and 0.0418 for P. clathratus and C. analis, respectively);
phylogeographic analysis found shallow topologies indicative of panmixia. These patterns
are similar to results found with other high gene-flow species. It is argued that Punta
Eugenia is only a barrier for strongly structured species that are equally isolated by smaller
points and distance.
Introduction
Geographic barriers can isolate marine populations and alter the effects of drift, selection
and gene flow (Slatkin, 1987). Examples include the Florida peninsula, USA (Bowen and
Avise, 1990), Cape Mendocino, California, USA (Cope, 2004) and the Baja peninsula, MX
(Stepien et al., 2001, Bernardi et al., 2003). Ocean currents (Shaw et al., 2004, Rocha-
Olivares and Vetter, 1999), distance (Alvarado-Bremer et al., 1996) and upwelling (Ayers
and Water, 2005) can also function as boundaries. The limited migration realized across
these breaks is important to management in that isolated populations will have a reduced
ability to adjust to extractive pressure. In an early example, Bowen and Avise (1990),
showed that fishing pressure on Centropristis striata (black sea bass) in the Gulf of Mexico
would not influence nor be ameliorated by populations on the Atlantic side of Florida.
51
Examples of isolation contrast the conventional wisdom that many marine species have high
gene flow (Ward et al., 1994). Unstructured species, if assayed around a boundary, would
modify interpretation of the barrier; the afore-mentioned sea bass results for example, were
one part of a project that surveyed two other species (Brevoortia tyrannus [menhaden] and
Acipenser oxyrinchus [sturgeon]), neither of which were as starkly influenced by the Gulf of
Mexico-Atlantic Ocean divide.
Individual studies may therefore show specific limitations to dispersal, but also skew
perceptions of broader patterns around a break. After initially encountering evidence that
Point Conception, California (see Figure 6) was both a biogeographic and phylogeographic
boundary, increased sampling resolution and a review of other species encouraged Burton
(1998) to conclude that Point Conception was not unique. In analyzing population structure
of a supralittoral marine copepod, Tigriopus californicus, Burton found other boundaries of
equal or greater magnitude that had never been identified as points of interest. While his
results did not change the fact that T. californicus was in some way limited across Point
Conception, he rejected the uniqueness of the boundary by finding that it only functions as a
phylogenetic break for species with populations already prone to isolation.
Other processes such as vicariance, rare oceanic events and life history patterns also
confound interpretations (e.g., Bernardi, 2000), further making it difficult to determine
conclusively if a location is a universal barrier to a diversity of species. This difficulty
increases the threshold amount of science needed to justify management decisions.
Institution of spatially complex management around a boundary would need to be species
specific for example, a noisome prospect in marine reserve design.
52
Figure 6. Map of sampling locations as described in Table 10.
53
One method for solidifying conclusions about a boundary is an assay of multiple species,
such as performed with disjunct taxa in the Gulf of California and the eastern Pacific
(Bernardi et al., 2003). That study also surveyed Point Conception, California (Burton,
1998) and Punta Eugenia, Baja, Mexico (Bernardi, 2000), two biogeographic breaks that are
potential (if not realized) intraspecific boundaries. While neither are associated with
vicariance hypotheses, both locations have upwelling patterns that could disrupt larval
settlement and reproductively isolate populations.
Genetic surveys across Punta Eugenia have been reported for ten species (Table 9), seven of
which have extremely limited dispersal across the point. The results for one species are
equivocal, and two indicate no evidence of a barrier. Like Point Conception and the Florida
peninsula, Punta Eugenia’s restriction of gene flow could be reflective of intrinsic species-
specific characteristics (as noted by Bernardi et al., 2003). Surveying more species across
Punta Eugenia will help clarify the limits to which the point can promote genetic structure.
We present genetic surveys of two species, the kelp bass (Paralabrax clathratus) and woolly
sculpin (Clinnocotus analis), whose ranges include Punta Eugenia. We worked to define
levels of genetic structure, compare our results to species already assayed across Punta
Eugenia, and build a more thorough understanding of the generalities that can be attributed
to Punta Eugenia as a phylogeographic boundary. Beyond testing for genetic breaks, this
research addressed management questions for the economically-important kelp bass, one of
the most heavily fished, sport resources in the State of California (Love, 1996).
54
Table 9. Record of species genetically surveyed across Punta Eugenia.
55
Kelp bass are abundant from the southern side of Punta Eugenia to Point Conception
(individuals are occasionally recorded as far south as Bahia Magdalena and north to the
Columbia River), a similar range to that of woolly sculpin (Eschmeyer et al., 1983). Both
species are part of the “Region 4” fauna of the Pacific coast (Point Conception to Punta
Baja) which frequently extends into “Region 3” (Punta Baja to Cabo San Lucas, Mexico)
(Parrish et al., 1981). Clinocottus analis is found in tide pools and Paralabrax clathratus is
a top predator in rocks and kelp of the near coast (Eschmeyer et al., 1983). The lack of
migrational tendencies by adult individuals in both species (Lowe et al., 2003, Yoshiyama et
al., 1992) and a strong homing ability in woolly sculpin (Yoshiyama et al., 1992) are
behaviors that could promote genetic structure. Larval stages can last a month for kelp bass
(Cordes and Allen, 1997) and sixty days for woolly sculpin (Davis and Levin, 2002) but
might be unable to move across oceanographic barriers.
Both species have been subject to genetic analysis with equivocal results. Yoshiyama and
Sassaman (1987) found no barriers with woolly sculpin sampled through northern and
central California, whereas Swank (1988), sampling from northern California to Lower Baja
California, noted an abrupt change in allele frequencies of the allozyme, Pgm I near Point
Conception, California (Figure 6). Hubbs (1926) used morphological characters of woolly
sculpin to distinguish two subspecies that he believed mixed around Los Angeles County
and Waples (1986) found structure among four distant sampling locations (La Jolla (San
Diego area), San Nicholas Island (see Figure 6), San Benito Island (off the tip of Punta
Eugenia) and Guadalupe Island (275 km west of the central Baja coast)) with 17
polymorphic loci. For kelp bass, Beckwitt (1983) and Waples (1987) found no genetic
structure, but Grothues (1994) found evidence of a north-south cline.
56
Materials and Methods
Sample Collection
One hundred and five kelp bass and 60 woolly sculpin were collected between June of 1999
and September of 2001 (Figure 6, Table 10). Paralabrax clathratus were caught with hook
and line using live bait (anchovy or sardine) and plastic lures. Fin tissue from each
individual was frozen in 1.7ml Eppendorf centrifuge tubes. Clinocottus analis were
collected with a dip net. After transport to the laboratory, fish were anesthetized using MS-
222 (250 mg / liter seawater) and either preserved in 70% EtOH or dissected.
Amplification and Sequencing
We isolated DNA with a tissue extraction kit from Gentra Biosystems ® was used to.
Polymerase chain reaction (PCR) amplifications were performed in a Genemate Genius
(Techne Inc.) with a protocol of 3 minutes: 93˚C, 35 cycles of 1 minute: 93˚C, 45 seconds:
54˚C and 1 minute 15 seconds: 72˚C, and then finally 5 minutes at 72˚C. The 25µl PCR
cocktails had 1µl of template and 2.5mM MgCl
2
, 2µM dNTPs, 1X Taq buffer, 0.5µM of
each primer and 0.5U of Taq DNA polymerase.
For the kelp bass, the entire mitochondrial control region (approximately 1100 base pairs)
was amplified using the primers 5!-AGCTCAGCGCCAGAGCGCCGGTCTTGTA-3! (“K”,
Lee et al., 1995) and 5!- GTCAGGACCATGCCTTTGTG (“12RH” Sivasundar et al.,
2001). An approximately 400 base pair fragment was amplified in woolly sculpin using the
primers “K” and “E“(5’-CCTGAAGTAGGAACCAGATG-3’(Lee et al., 1995), which is
located in an internal conserved region of the mitochondrial control region.
57
Table 10. Sample locations (marked with letters in Figure 6) with sample size (N), sample
dates, count of unique haplotypes in each population, haplotype diversity (h), sequence
diversity ("), mean pairwise differences (MPD) and Tajima’s neutrality statistic (D).
58
Samples were run on a 1% agarose gel, stained with Ethidium Bromide (EtBr) and
visualized on a UVP UV transilluminator. Bands were excised and purified with a Qiagen
Qiaquick® gel extraction kit. Sequencing was performed on either an ABI 373 or ABI 377
automated sequencer (Applied Biosystems). Kelp bass samples were sequenced with a
primer modified from “A” (Lee et al., 1995) (5!- TACCTCTACCCCCAAAGCTAGG-3!)
and woolly sculpin were sequenced with primer “E.” Sequences are deposited on Genbank;
accession numbers are DQ192295-DQ192399 for Paralabrax clathratus and EF077714-
EF077773 for Clinocottus analis.
Data Analysis
After alignment and trimming in Sequencher 4.5 (Gene Codes Corporation), sequences were
499 bases for Paralabrax clathratus (trimmed to 363 bases in the phylogenetic analysis to
include the the outgroups, Paralabrax maculatofasciatus and Paralabrax nebulifer) and 321
base pairs for Clinocottus analis (increased from 320 base pairs with the inclusion of
Oligocottus maculosus). Phylogeographic relationships were analyzed with MrBayes
(Huelsenbeck and Ronquist, 2001) using a general time reversible (GTR) model and gamma
distributed variation in substitution rate. Data from a number of potentially informative
insertion/deletions (indels) in the woolly sculpin sequences were preserved by including
them as a second, presence/absence data type. The program was run until the average
standard deviation of split frequencies between the chains had reached 0.01; kelp bass were
run for 20,000,000 generations (sample frequency=1/20,000) and woolly sculpin for
1,000,000 generations (sample frequency=1/1000). Consensus trees were generated with
removal of the first 25% of the samples and are reported with posterior probabilities.
59
Population structure was analyzed in ARLEQUIN 3.01 (Excoffier et al., 2005). Genetic
variation was described with estimates of nucleotide diversity (", Nei, 1987), haplotype
diversity (h, Nei, 1987) and mean pairwise difference between sequences. Neutrality was
tested with D (Tajima, 1989). Variance among populations and regions was partitioned by
analysis of molecular variance (AMOVA) according to five hypotheses: no regions, a break
at Punta Eugenia, a break between coastal and channel islands (kelp bass only), four
groupings (south of Punta Eugenia, north Baja peninsula, channel islands and Point
Conception area, kelp bass only), and a break at Point Conception (woolly sculpin only).
Pair-wise differences, Jukes and Cantor (Jukes and Cantor, 1969) and Kimura 2p (Kimura,
1980) corrections were used to set distances. Statistical confidence for all tests was
determined with 10,000 permutations (as described in (Excoffier et al., 1992)).
We assessed isolation by distance with a Mantel test (exact p-value calculated with 10,000
permutations) using distances between latitude/longitude coordinates calculated at:
http://jan.ucc.nau.edu/~cvm/latlongdist.html. Distances reflected the shortest possible
oceanic routes and were log-transformed for comparison to linearized F
ST
values (Slatkin,
1995) calculated in ARLEQUIN. Measurements were done in steps between sample
location coordinates (Table 10) and Punta Eugenia (27.86°N, 115.07°W), Punta San Antonio
(29.75°N, 115.75°W), the eastern tip of Santa Rosa Island (33.97°N, 119.95°W), and Point
Conception (34.45°N, 120.70°W). Using coastline distances would have been problematic
with Channel Island populations; straight distances on the other hand, were unreasonable for
some sample pairs, such as Abreojos and Morro Santo Domingo (Locations A and D, Figure
6), which are located on opposite sides Punta Eugenia.
60
Results
Paralabrax clathratus had a large number of haplotypes (53 alleles in the 105 samples) with
40 that were unique (found in only one individual). Haplotype diversity was high but
nucleotide diversities and mean pairwise differences indicated subtle variation between
haplotypes (Table 10). Pairwise differences between haplotypes averaged three substitutions
(~1%); 19 were one base pair different than the common haplotype and 42 were three or
less. A single haplotype had the maximum difference of eight substitutions. Clinocottus
analis had lower haplotype diversity but also exhibited shallow haplotype differences (Table
10). Eight haplotypes (two unique) had a mean pairwise difference of 0.9 substitutions with
a maximum difference of only one. Most of the haplotype diversity was a series indel events
around a five base pair G-tract, which defined five of the seven haplotypes.
The significant results from the neutrality tests (Tajima’s D, Table 10) enumerated these
patterns of shallow diversity in both species. For both species, haplotype diversity was high
in relationship to sequence diversity, a pattern attributed to a recent expansion in population
size (Grant and Bowen, 1998) or a series of bottlenecks (Hedgecock, 1994).
The shallowness of the species’ diversities was also seen in the phylogenies (Figures 7 and
8). Branching did not exceed three steps for kelp bass and one step for woolly sculpin. The
star-like phylogenies showed no evidence of a break at Punta Eugenia or any other location
for the two species. A common allele was found in nine out of ten populations for the kelp
bass and six out of six for the woolly sculpin. The few clades that appear out of the cluster
of common alleles contained a mix of individuals from both proximal and distant
populations with no apparent geographic organization.
61
Figure 7. Bayesian consensus phylogeny for 363 bases of the mitochondrial control region
in Paralabrax clathratus with outgroups of P. maculatofasciatus and P. nebulifer. Posterior
probability values greater than 50% are shown below nodes; location identifiers are used as
described in Table 10 and Figure 6.
62
Figure 8. Bayesian consensus phylogeny for 321 bases of the mitochondrial control region in
Clinocottus analis with an outgroup of Oligocottus maculosus. Posterior probability values
greater than 50% are shown below nodes; location identifiers are used as described in Table
10 and Figure 6.
63
This pattern was reflected in the results of the AMOVA analyses (Table 11); only the
northern break hypothesis for woolly sculpin was statistically significant. Samples from the
Pigeon Point collection site (the most northern) found two unique alleles, but with a p-value
of 0.048, this lone significant result did not survive a Bonferroni correction.
Pairwise population #
ST
values averaged 0.0418 for Clinocottus analis and 0.0144 for
Paralabrax clathratus. Both values are typical of the marine environment (Ward et al.,
1994), suggesting near panmixia across populations. The number of pairwise differences
between each individual were so few that the Jukes and Cantor and Kimura 2P distance
corrections did not change the significance of any of the results.
Although the AMOVA tests were non-significant, the pairwise linear F
ST
values were tested
for an isolation by distance signal. This signal was not seen and the conclusion of panmixia
was furthered bolstered by non-significant results from Mantel Tests (r=-0.157, p=0.854 for
kelp bass, r=0.379, p=0.058 for woolly sculpin).
Discussion
Population Structure and Punta Eugenia Break
These results indicated that panmixia is the most reasonable interpretation of the data. There
was no break at Punta Eugenia and only a mild suggestion of subdivision between southern
and northern samples of woolly sculpin. While rare or low frequency alleles can appear
isolated, a mix of locations contributing to substructure in haplotype phylogenies and the
ubiquity of common alleles indicated high gene flow. A conclusion of no structure in
woolly sculpin contrasts the work of Waples (1987) and Swank (1988). In the case of the
64
Table 11. Results from AMOVA tests with five hypotheses: (1) no regions, (2) southern
region of locations A, B and C and northern region of all other samples, (3) island region
with populations G, H and I and coastal region with all other samples, (4) four regions: A, B
and C; D and F; G, H and I; L and M, and (5) a northern region of N and O and a southern
region of all other samples. *p<0.05.
65
Waples (1987) result, sampling from the isolated islands of San Nicholas, Guadalupe and
San Benitos might find higher genetic structure then our sampling along the continental
coast because Clinocottus analis larvae are typically caught near shorelines (Waples, 1987).
Increased sampling resolution might resolve this question.
A sensitivity analysis of sampling power following the methodology of Vogel and Edmands
(in prep) found that the average per-site sample sizes of these studies could have registered
($=0.2) statistically significant results for populations with #
ST
values greater than 0.125 for
Paralabrax clathratus and 0.254 for Clinocottus analis. While these values are within the
range found by Bernardi et al. (2003), they are high for marine species (e.g., Ward et al.,
1994), suggesting that sample size was sufficient only to detect significant structure in
highly subdivided populations.
Despite the limited sampling power, there is little doubt that gene flow for the two species is
high. First, while F
ST
values of low-structure populations are prone to imprecision and
inaccuracy (Waples, 1998), they still provide a reasonable estimate. When sampling ten
individuals from each of two simulated populations (as in Vogel and Edmands, in prep) of
kelp bass with a #
ST
value of 0.015, 85% of the samples had #
ST
estimates below 0.05 and
97% below 0.1. For woolly sculpin, ten samples out of each population with a #
ST
value of
0.04 recorded a result below 0.1 83% of the time. So while statistical significance cannot be
attained with the sample sizes used in this study, it would have taken single location sample
sizes of about 90 for kelp bass [#
ST
=0.0144] and 75 for woolly sculpin [#
ST
=0.0418]) and it
is doubtful that “true” F
ST
values are high.
66
Secondly, Paralabrax clathratus and Clinocottus analis both had a ubiquitous common
allele and multiple populations sharing phylogenetic diversity. Increased sampling would
have been a search for statistically significant allele frequency differences and not
establishment of location-specific compliments of haplotypes. Without high-frequency,
localized alleles, true F
ST
values are certainly low (e.g., Slatkin, 1987).
Thirdly, sample number for these two studies is well within the range of the other ten studies
listed in Table 9. While not perfect, it allows an equivalent resolution in looking at Punta
Eugenia as a barrier. Those studies that found structure across Punta Eugenia had sufficient
sampling to find a phylogenetic signal, something we clearly did not find.
If kelp bass and woolly sculpin truly have barrier-isolated populations, sample sizes on the
order of 100-150 individuals will be required to find statistical significance. Because these
two species have such low pairwise #
ST
values (and high migration rates), temporal
sampling would be required to confirm true structure (as noted by Waples, 1998). Further,
comparisons between populations north and south of Punta Eugenia would have to show
higher gene flow limitation than other pairwise calculations. As can be seen in the #
CT
values in Table 11, this is a pattern that does not seem probable.
One possible population history would be recent isolation of populations having such large
effective sizes that divergence is immeasurably slow. In this case, isolation would have
recently occurred, as inferred from the shallow phylogenies, and would require a complex
argument involving a population explosion, range advance and behavior modification to
limit gene flow. Panmixia is a simpler explanation for both kelp bass and woolly sculpin.
67
Comparison to other Species
There is a dichotomy in the patterns of structure across Punta Eugenia. Species show no
limitation at Punta Eugenia and are unstructured in general, or are limited at Punta Eugenia
and show high levels of isolation across many if not all sample locations. The panmictic
patterns found with Paralabrax clathratus and Clinocottus analis in this study are
reminiscent of those noted by Bernardi et al. (2003) for rock wrasse (Halichoeres
semicinctus, mean F
ST
=0.01) and California sheephead (Semicossyphus pulcher, mean
F
ST
=0.00). Gene flow for these four species appears to be high between all locations.
On the other hand are the eight species that show a genetic break around Punta Eugenia
(Table 9). Because of the dichotomy in structuring, limitation might be more reflective of
intrinsic factors than Punta Eugenia itself (Bernardi et al., 2003). With Girella nigricans for
example, F
ST
values could be inflated by the presence of a cryptic species (Terry et al.,
2000). Girella nigricans was historically considered to be two species (G. nigricans and G.
simplicidens) and has some of the highest sequence diversities (3.4-4.9% between clades)
found in marine fish (Terry et al., 2000). An alternative explanation that would align G.
nigricans’ population structure with its dispersal ability would be reflective of Avise's
(2000) category II phylogenetic pattern. This species might have high gene flow along the
Pacific coast but have southern populations infused with gene lineages that allopatrically
diverged in the Sea of Cortez.
In the case of Embiotoca jacksoni, the species does not have a pelagic larval stage (Huang
and Bernardi, 2001). Huang and Bernardi (2001) also found evidence of other geographic
breaks across Santa Monica Bay near Los Angeles and Big Sur/Morro Bay. With this
68
overall, high level of structure, the question arises as to whether species that provide
evidence that Punta Eugenia is a phylogeographic boundary are regularly isolated, such as
what was found with Tigriopus californicus across Point Conception (Burton, 1998).
Regular isolation would indicate that there is nothing unique about Punta Eugenia per se, but
that population structure is a reflection of general dispersal inabilities.
In addition to Embiotoca jacksoni, five of the species that are isolated across Punta Eugenia
(Table 9) are in fact highly structured. Along the Pacific coast, every pairwise comparison is
significant with the longjaw mudsucker (Gillichthys mirabilis, #
ST
=0.448) (Huang and
Bernardi, 2001) and the California killifish (Fundulus parvipinnis, #
ST
=0.844) (Bernardi and
Talley, 2000) using pairwise comparisons we calculated with referenced Genbank sequences
and ARLEQUIN 3.01. We also analyzed Restriction Fragment Length Polymorphism
(RFLP) data from the California halibut (Paralichthys californicus) (Morgan, 1997) and
found significant structure (#
ST
=0.720) for all pairwise sample locations except one (Santa
Catalina Island-San Diego Bay). For the red spiny lobster (Panulirus interruptus), only two
of ten comparisons were non-significant (Abreojos-Punta Eugenia [see Figure 6] and Bahía
Magdalena-Ensenada) using allozymes (overall F
ST
=0.101) (Perez-Enriquez et al., 2001).
Analyses of the fifth species, spotted sand bass (Paralabrax maculatofasciatus), have been
equivocal. Bernardi et al. (2003) found restricted gene flow across Punta Eugenia using data
from Tranah and Allen (1999) and Stepien et al. (2001), but neither study individually
demonstrated a barrier. Stepien et al. (2001) did not sample around Punta Eugenia (there are
no Pacific coast sample sites south of Punta Eugenia) and Tranah and Allen (1999) explicitly
state that the sample sites of Guerrero Negro (located in the throat of Bahia de Vizscaíno and
69
north of Punta Eugenia, see Figure 6) and Magdalena Bay (south of Punta Eugenia) were the
most closely related locations in the study. A later analysis by Salomon (2005) did find
significant structure (global #
ST
=0.916). The last species to show isolation, the pargo
(Anisotremus davidsonii) (Bernardi et al., 2003), cannot be further analyzed because of low
sample size (n=26). In general though, species sampled across Punta Eugenia show no
population structure, or near complete differentiation across all sample locations.
Another line of support for the theory that Punta Eugenia is an important boundary is the fact
that in three of these differentiated species, sample sites south of Punta Eugenia were
particularly divergent. However, other factors such as small populations at the species’
range limit (Bernardi, 2000), small effective sizes (Bernardi and Talley, 2000) and cryptic
subspecies (Bernardi and Talley, 2000) could be unevenly increasing F
ST
values.
We explored this idea further with data from Fundulus parvipinnis, Embiotoca jacksoni and
Gillichthys mirabilis, the three species which show this pattern. Pairwise F
ST
differences
were all high; comparisons across Punta Eugenia averaged 0.948, 0.637 and 0.732 for the
three species respectively, as opposed to 0.577, 0.530 and 0.138 between locations that do
not span the point (Bernardi and Talley, 2000, Bernardi, 2000, Huang and Bernardi, 2001).
We tested for isolation by distance using Mantel tests (as described above) of published
genetic data and geographic distances we calculated (Table 12). Were there this isolation,
values across Punta Eugenia would be increased because of distance, and not completely
indicative of the barrier strength of the point. For F. parvipinnis, we calculated a significant
signal (r=0.766, p=0.001). We also find isolation by distance in E. jacksoni (Bernardi, 2000)
(r=0.646, p=0.0003) and G. mirabilis (Huang and Bernardi, 2001) (r=0.697, p=0.048).
70
Table 12. Sampling coordinates used for isolation by distance analysis of data from
Fundulus parvipinnis (Bernardi & Talley 2000), Embiotica jacksoni (Bernardi 2000) and
Gillichthys mirabilis (Huang & Bernardi 2001).
71
Implications
We are not convinced that Punta Eugenia is any more of a phylogeographic barrier than
Point Conception. Like Tigriopus californicus across Point Conception (Burton, 1998), the
evidence is dependent on species that show high levels of isolation. Were a species found to
be isolated by Punta Eugenia but generally panmictic within northern and southern
populations, a more definitive declaration might be made. There are to date no published
analyses of genetic structure that can fit this model and more thoroughly confirm that Punta
Eugenia is both a biogeographic and phylogeographic boundary. While Punta Eugenia
seems to isolate populations more often than Point Conception, only species with generally
low levels of gene flow appear excessively prone to restrictions in dispersal around the point.
Kelp bass and woolly sculpin have high levels of migration. As long as robust population
sizes are maintained, managing kelp bass as one panmictic stock is in agreement with the
genetic structure. Homogenization can occur with low migration levels (Wright, 1931),
these results do not preclude the possibility of localized extinction due to fishing pressure, or
perhaps severe storms in the case of woolly sculpin. Assuming no demographic constraints,
the genetic structure reported here indicates that these types of events should ultimately, if
not immediately, be mitigated by re-colonization from other locations.
Acknowledgements
This research was supported by a Boren Fellowship, the Wrigley Institute of Environmental
Studies, NSF grant DBI-9904694 and USC’s Zumberge Faculty Research and Innovation
Fund. We thank H. Kochounian for advice with sequencing primer design, and J. Flowers
and J. Diamond for support with collection of samples.
72
Chapter 2 References
ALVARADO-BREMER, J. R. A., MEJUTO, J., GREIG, T. W. & ELY, B. (1996) Global
population structure of the swordfish (Xiphias gladius L.) as revealed by analysis of
the mitochondrial DNA control region. Journal of Experimental Marine Biology and
Ecology, 197, 295-310.
AVISE, J. (2000) Phylogeography: the history and formation of species, Cambridge, MA,
Harvard University Press.
AYERS, K. & WATER, J. (2005) Marine biogeographic disjunction in central New Zealand.
Marine Biology, 147, 1045-1052.
BECKWITT, R. (1983) Genetic Structure of Genyonemus lineatus, Seriphus politus
(Sciaenidae) and Paralabrax clathratus (Serranidae) in Southern California. Copeia,
3, 691-696.
BERNARDI, G. (2000) Barriers to gene flow in Embiotoca jacksoni, a marine fish lacking a
pelagic larval stage. Evolution, 54, 226-237.
BERNARDI, G., FINDLEY, L. & ROCHA-OLIVARES, A. (2003) Vicariance and dispersal
across Baja California in disjunct marine fish populations. Evolution, 57, 1599-1609.
BERNARDI, G. & TALLEY, D. (2000) Genetic evidence for limited dispersal in the coastal
California killifish, Fundulus parvipinnis. Journal of Experimental Marine Biology
and Ecology, 255, 187-199.
BOWEN, B. W. & AVISE, J. C. (1990) Genetic structure of Atlantic and Gulf of Mexico
populations of sea bass, menhaden, and sturgeon: influences of zoogeographic
factors and life-history patterns. Marine Biology, 107, 371-381.
BURTON, R. (1998) Intraspecific phylogeography across the Point Conception
biogeographic boundary. Evolution, 52, 734-745.
COPE, J. (2004) Population genetics and phylogeography of the blue rockfish (Sebastes
mystinus) from Washington to California. Canadian Journal of Fisheries and
Aquatic Science, 61, 332-342.
CORDES, J. F. & ALLEN, L. G. (1997) Estimates of age, growth, and settlement from
otoliths of young-of-the-year Kelp Bass (Paralabrax clathratus). Bulletin of the
Southern California Academy of Sciences, 96, 43-60.
DAVIS, J. & LEVIN, L. (2002) Importance of pre-recruitment life-history stages to
population dynamics of the woolly sculpin, Clinocottus analis. Marine Ecology
Progress Series, 234, 229-246.
ESCHMEYER, W. N., HERALD, E. S. & HAMMANN, H. (1983) A Field Guide to Pacific
Coast Fishes, New York, New York, Houghton Mifflin Company.
EXCOFFIER, L., LAVAL, L. & SCHNEIDER, S. (2005) Arlequin ver. 3.0: An integrated
software package for population genetics data analysis. Evolutionary Bioinformatics
Online, 1, 47-50.
EXCOFFIER, L., SMOUSE, P. & QUATTRO, J. (1992) Analysis of molecular variance
inferred from metric distances among DNA haplotypes: application to human
mitochondrial DNA restriction data. Genetics, 131, 479-491.
GRANT, W. & BOWEN, B. (1998) Shallow population histories in deep evolutionary
lineages of marine fishes: insights from sardines and anchovies and lessons for
conservation. The Journal of Heredity, 89, 415-426.
GROTHUES, T. M. (1994) An investigation into the phenetic structure and larval dispersal
patterns of the kelp bass, Paralabrax clathratus. Marine Biology. Los Angeles,
California State University, Northridge.
73
HEDGECOCK, D. (1994) Does variance in reproductive success limit effective population
sizes of marine organisms? IN BEAUMONT, A. (Ed.) Genetics and Evolution of
Aquatic Organisms. London, Chapman and Hall.
HUANG, D. & BERNARDI, G. (2001) Disjunct Sea of Cortez-Pacific Ocean Gillychthys
mirabilis populations and the evolutionary origin of the their Sea of Cortez endemic
relative. Marine Biology, 138.
HUBBS, C. (1926) Descriptions of new genera of cottoid fishes related to Artedius.
Occasional Papers of the Museum of Zoology, University of Michigan, 1-16.
HUELSENBECK, J. & RONQUIST, F. (2001) MRBAYES: Bayesian inference of
phylogeny. Bioinformatics, 17, 754-755.
JUKES, T. & CANTOR, C. (1969) Evolution of protein molecules. IN MUNRO, H. (Ed.)
Mammalian Protein Metabolism. New York, Academic Press.
KIMURA, M. (1980) A simple method for estimating evolutionary rate of of base
substitution through comparative studies of nucleotide sequences. Journal of
Molecular Evolution, 16, 111-120.
LEE, W., CONROY, J., HOWELL, W. & KOCHER, T. (1995) Structure and function of
teleost mitochondrial control regions. Journal of Molecular Evolution, 41, 54-66.
LOVE, M. (1996) Probably More than You Want to Know About the Fishes of the Pacific
Coast, Santa Barbara, CA, Really Big Press.
LOWE, C., TOPPING, D., CARTAMIL, D. & PAPASTAMATIOU, Y. (2003) Movement
patterns, home range, and habitat utilization of adult kelp bass Paralabrax clathratus
in a temperate no-take marine reserve. Marine Ecology Progress Series, 256, 205-
216.
MORGAN, M. (1997) Genetic variation among nine populations of California halibut,
Paralichthys californicus. Biology. Northridge, CA, California State University,
Northridge.
NEI, M. (1987) Molecular Evolutionary Genetics, New York, Columbia University Press.
PARRISH, R., NELSON, C. & BAKUN, A. (1981) Transport mechanisms and reproductive
success of fishes in the California Current. Biological Oceanography, 1, 175-203.
PEREZ-ENRIQUEZ, R., VEGA, A., AVILA, S. & SANDOVAL, J. (2001) Population
genetics of red spiny lobster (Panulirus interruptus) along the Baja California
peninsula, Mexico. Marine and Freshwater Research, 52, 1541-1549.
ROCHA-OLIVARES, A. & VETTER, R. D. (1999) Effects of oceanographic circulation on
the gene flow, genetic structure, and phyogeography of the rosethorn rockfish
(Sebastes helvomaculatus). Canadian Journal of Fisheries and Aquatic Science, 56,
803-813.
SALOMON, M. (2005) Gene flow among populations of the spotted sand bass, Paralabrax
maculatofasciatus, from southern California and two populations along the pacific
coast of Baja, California. Biology. Northridge, CA, California State University,
Northridge.
SHAW, P., ARKHIPKIN, A. & AL-KHAIRULLA, H. (2004) Genetic structuring of
Patagonian toothfish populations in the Southwest Atlantic Ocean: the effect of the
Antarctic Polar Front and deep-water troughs as barriers to genetic exchange.
Molecular Ecology, 13, 3293-3303.
SIVASUNDAR, A., BERMINGHAM, E. & ORTÍ, G. (2001) Population structure and
biogeography of migratory freshwater fishes (Prochilodus: Characiformes) in major
South American rivers. Molecular Ecology, 10, 407-418.
74
SLATKIN, M. (1987) Gene flow and the geographic structure of natural populations.
Science, 236, 787-792.
SLATKIN, M. (1995) A measure of population subdivision based on microsatellite allele
frequencies. Genetics, 139, 457-462.
STEPIEN, C., ROSENBLATTT, R. & BARGMEYER, B. (2001) Phylogeography of the
spotted sand bass, Paralabrax maculatorfasciatus: divergence of Gulf of California
and Pacific Coast populations. Evolution, 55, 1852-1862.
SWANK, S. (1988) Biochemical systematics of the genus Clinocottus (Cottidae). Bulletin of
the Southern California Academy of Science, 87, 57-66.
TAJIMA, F. (1989) Statistical method for testing the neutral mutation hypothesis by DNA
polymorphism. Genetics, 123, 585-595.
TERRY, A., BUCCIARELLI, G. & BERNARDI, G. (2000) Restricted gene flow and
incipient speciation in disjunct Pacific Ocean and Sea of Cortez populations of a reef
fish species, Girella nigricans. Evolution, 54, 652-659.
TRANAH, G. & ALLEN, L. (1999) Morphological and genetic variation among six
populations of the spotted sand bass, Paralabrax maculatofasciatus, form southern
California to the upper Sea of Cortez. Bulletin of the Southern California Academy
of Sciences, 98, 103-118.
VOGEL, A. & EDMANDS, S. (in prep) The influence of sample size and population
differentiation on statistical power and estimation of F
ST
in marine species.
WAPLES, R. (1986) A multispecies approach to the analysis of gene flow in marine shore
fishes. Marine Biology. University of California, San Diego.
WAPLES, R. (1987) A multispecies approach to the analysis of gene flow in marine shore
fishes. Evolution, 41, 385-400.
WAPLES, R. (1998) Seperating the wheat from the chaff: patterns of genetic differentiation
in high gene flow species. Journal of Heredity, 98, 438-450.
WARD, R., WOODWARK, M. & SKIBINSKI, D. (1994) A comparison of genetic diversity
levels in marine, freshwater and anadromous fish. Journal of Fish Biology, 44, 213-
232.
WRIGHT, S. (1931) Evolution in Mendelian populations. Genetics, 16, 97-159.
YOSHIYAMA, R., GAYLORD, K., PHILIPPART, M., MOORE, T., JORDAN, J., COON,
C., SCHALK, L., VALPEY, C. & TOSQUES, I. (1992) Homing behavior and site
fidelity in intertidal sculpins (Pisces: Cottidae). Journal of Experimental Marine
Biology and Ecology, 160, 115-130.
YOSHIYAMA, R. & SASSAMAN, C. (1987) Geographical patterns of allozymic variation
in three species of intertidal sculpins. Environmental Biology of Fishes, 20, 203-218.
75
Chapter 3: Development and inheritance of molecular markers in the kelp bass,
Paralabrax clathratus
Abstract
Although genetic analysis is useful to fishery managers, it is expensive, can necessitate
lengthy optimization, and require confirmation of Mendelian inheritance in the marker set.
We report a cost effective method to assay polymorphism in the mitochondrial control
region and confirm Mendelian inheritance of eight microsatellite markers in the kelp bass,
Paralabrax clathratus. A set of restriction enzymes were found to give equivalent patterns
of geographic distance when compared to sequence data, and all microsatellites tested show
Mendelian inheritance. These tools present the opportunity to study more thoroughly the
genetic structure of P. clathratus, and assist in its management as a fisheries species.
Introduction
Genetic analysis can strengthen understanding of fishery species (Ward, 2000) through
delineation of stocks (Graves and McDowell, 2003), description of gene flow (Laurent et al.,
2006), and discovery of cryptic species (Rocha-Olivares et al., 1999). With this type of
information, managers are better able to set catch limits, define protected areas and instigate
other measures for protecting the health of fishery populations.
While molecular tools have transformed fisheries biology, their development and
implementation can be expensive and require extensive optimization (Burton, 1996). This
process is especially costly for sequencing; commercial services cost ~ $10 sample
(http://www.genomex.com, http://www.laragen.com) for 500-700 bases of PCR product, and
do not include the price of purifying templates. Counting only consumables and ignoring the
76
sequencer purchase price, the service contract and any labor and trouble shooting, we have
calculated that in-house sequencing of 500-700 bases on a Beckman CEQ 8000 costs ideally
~$4-5 per sample. Loci scored as fragments on an automated sequencer are also expensive
(Schuelke, 2000) and require development of a library, and involve extensive preparation
with bacterial colonies and cost commercially ~US$12,500 (Morin et al., 2004).
Another cost can be accrued in confirming Mendelian inheritance of the marker set.
Although Mendelian inheritance is rarely tested, a recent survey found that one in fifteen
microsatellite loci had non-Mendelian patterns (Selkoe and Toonen, 2006). Departure from
expected Mendelian ratios can occur because of polyploidy and/or gene duplication (Baums
et al., 2005), sex linkage (Van't Hof et al., 2005), null alleles (Selkoe and Toonen, 2006),
genetic imprinting (de Meeus et al., 2004), non-specific amplification (Smith et al., 2000),
and a host of complex and poorly understood gene-gene and environment-gene interactions
(van Heyningen and Yeyati, 2004). Recognizing these factors is important because evidence
of non-Mendelian inheritance would compromise any statistical test of the genetic data.
The reporting of cost-effective methodologies and confirmational studies of inheritance
patterns can therefore be invaluable in the initiation of a research study. A set of markers
will be available for not only the species in question, but frequently for its congeners as well.
According to this rationale, this communication presents genetic methods developed for the
kelp bass, Paralabrax clathratus.
Kelp bass are a temperate serranid found abundantly from Point Conception to just south of
Punta Eugenia, Mexico (Eschmeyer et al., 1983), and are an important component of the
77
$300 million dollar California sport fishing industry (Love, 1996, Thomson, 2001). The
species is a ubiquitous predator (Anderson, 2001) in kelp and rock communities.
Paralabrax clathratus occurs in current and proposed southern California marine reserves
and has been the subject of a diversity of research (Anderson, 2001, Carr, 1994, Lowe and
Topping, 2003, Waples and Rosenblatt, 1987).
To assist research efforts related to this species, the functionality of restriction fragment
length polymorphism (RFLP) analysis of the mitochondrial control region is documented.
Mitochondrial markers have been extensively used in population biology (e.g., Rocha-
Olivares and Vetter, 1999, Cope, 2004), but frequently require costly sequencing protocols
as noted above. A population assay with RFLPs on the other hand, can assess a substantial
amount of the sequence diversity for a fraction of the cost. We also present the results of a
cross to confirm Mendelian inheritance of all microsatellite markers in Selkoe et al. (2005).
Although difficult to perform with many marine fish (Takagi et al., 2003), we successfully
implemented a series of controlled crosses and scored both parental and filial genotypes.
Materials and Methods
Development of RFLPs
Sequences for the mitochondrial control region of 105 kelp bass (Genbank accession
numbers DQ192295-DQ192399) from 10 locations (Vogel and Edmands, in prep-a) were
used in the analysis. Alignments were performed in Sequencher 4.5 (Gene Codes
Corporation) and sequences surveyed with the program’s restriction enzyme library.
Sequence fragments of 40 base pairs (twenty to each side of the polymorphism) were created
and tested against the 280 enzyme options of the program.
78
Pairwise !
ST
values for the 10 locations were calculated with Arlequin 3.01 (Excoffier et al.,
2005). Source sequences were trimmed to 499 base pairs so as to include all individuals,
after which population structure was analyzed for both sequence and RFLP data. The level
of correlation between the two methods was calculated with a Mantel test.
Inheritance Patterns
Microsatellites from (Selkoe et al., 2005) were tested for Mendelian inheritance at the
Wrigley Institute for Environmental Studies on Santa Catalina Island, California in August
of 2005. An improved reverse primer (5"-TGTATTTAACTTTTAGCCCC-3") was created
for AV115, which appeared prone to stutter patterns. A new microsatellite that was originally
isolated in Polyprion americanus by M. Zatcoff and R. Chapman, Pam13 (Genbank
accession number DQ845458), was also tested with the primers: 5"-
CGCATGTTTGTAAGAACAGGAAG-3" and 5"- CCACTCACTGGTGCAGAAAC-3".
Four males and four females (1.5 to 7 kg) were caught with hook and line and held in 1000
gallon running seawater tanks. The fish were fed anchovy and squid and given PVC piping
to provide shelter and reduce stress. Sexing was performed with 1.5 mm cannulation tubing.
After anesthetization with 20 ppm Quinaldine, each individual was injected intramuscularly
with 20 µg/kg body weight of luteinizing hormone releasing hormone analog (LHRH-a).
After 24 hours, fish were re-injected with the same concentration of LHRH-a and monitored
for signs of spawning behavior and hydrated and clear eggs. Courtship by males of highly
gravid females was evident 26 hours after the second injection, at which point most eggs
were at least 1000 µm and had a taut, round appearance.
79
Two males and two females were stripped, and eggs and milt mixed in two gallons of
seawater for five minutes. Fertilized eggs were added to bubbling, 15 gallon tanks and
harvested after 12 hours (development to the blastula stage and beyond was checked by
microscope) for immediate DNA extraction with phenol chloroform.
Mendelian segregation was checked with contingency tests of genotype counts. Evidence
for linkage was tested with a natural log test of r (using a #
2
distribution to ascertain
significance) for each combination of microsatellite loci, excluding those that involved a
homozygous genotype in one of the parents.
Results and Discussion
Development of RFLPs
Seventy-seven enzymes diagnosed 34 of the 51 polymorphisms in the control region marker
(Table 13). The enzymes AciI, DdeI, !TaqI, and SfaNI (Table 13), which distinguish eleven
of the substitutions, were used to compare genetic structure as determined by RFLP or
sequence scores. Digestions of the 1000 base pair band (primers “K” and “12RH,” protocol
from Vogel and Edmands, in prep-a), with a mixture of AciI and DdeI were scored on a
Spreadex acrylamide gel (EL1200, Elchrom Scientific, http://www.elchrom.com). A
separate digestion with !TaqI was run on a denser gel, the EL400.
The most abundant polymorphism (position 233) was cut by SfaNI (Table 13). Because of
enzyme cost, a new primer (233C, 5"-CGATTAGGCGTGKCGACATG-3") was designed to
recognize the 3" edge of the variable site. The 18
th
base of the primer does not match the
80
Table 13. Restriction enzymes that recognize polymorphism in the mitochondrial control
region of Paralabrax clathratus. Polymorphic nucleotide positions correspond to heavy
chain sequences of accession numbers: DQ192295-192399. Annotated enzymes have the
following isoschizomers:
1
AvaII, SinI;
2
BsiHKAI, HgiAI;
3
Bsp1286I, SduI;
4
Alw44I, SnoI;
5
MvaI;
6
Cfr13I, Sau96I;
7
AquI, BcoI;
8
MaeI, RmaI, XspI;
9
Ce11II, EspI;
10
SecI;
11
BstOI,
EcoRII;
12
MflI, XhoII;
13
EcoO109I, PssI;
14
Tsp509I;
15
HindII;
16
MaeII;
17
BfuCI, Sau3AI,
NdeII, DpnI, DpnII;
18
Tru9I;
19
TthHB8I. **Polymorphism recognized by enzyme SfaNI can
be recognized with AflIII if amplified with primers “K” and “233cutter.”
81
Table 13. Continued.
82
Table 13. Continued.
83
strand sequence and creates a change (233 is underlined) from 5"-GCATCTAC-3" to 5"-
GCATGTAC-3". A cheaper enzyme, AflIII, can then be used to assay Adenine substitutions
at 233. The digested product (amplified with “K” and the same protocol in (Vogel and
Edmands, in prep-a)) can be double-loaded in a gel with AciI and DdeI.
Pairwise differences (!
ST
) of the two data types (Table 14) were significantly correlated
(r=0.6551, p<0.0001) (Figure 9). Because reduced haplotype diversity in RFLP data (0.758
from 0.9081) lowered within-location distances (analogous to H
S
), the global !
ST
value was
higher with RFLP data (0.0374) than sequence data (0.0144). Four of forty-five pairwise
comparisons with sequence data and 6/45 with RFLP data were significant (Table 14).
Differences in both analyses were connected to a single population; the two extra significant
results in the RFLP data reinforced the pattern of the sequence data.
The RFLP protocol cost $2-3 per sample to score PCR product, about half the price of in-
house sequencing, and only required a thermal cycler and a 600ml gel box, as opposed to an
automated sequencer. While information is always lost in an RFLP analysis as compared to
sequencing, the robustness of this method with Paralabrax clathratus allows for other cost-
effective analyses of population structure (e.g., Vogel and Edmands, in prep-b).
Inheritance Patterns
Eight microsatellite loci were amplified from offspring and parents of three crosses (Table
15). One anticipated genotype was not found in a case where only 8 progeny were scored
(microsatellite AV6, cross 6 $ 4, genotype 197217). Crosses produced no unexpected
genotypes, as might be expected were null alleles segregating or amplifications nonspecific.
84
Table 14. Comparison between sequence data and RFLP scores for one hundred and five
individuals divided across ten populations: TR=Tahiguas Reef, NR=Naples Reef, SN=San
Nicholas Island, SC=Santa Catalina Island, CE=San Clemente Island, SQ=San Quintín,
MSD=Morro Santo Domingo, BT=Bahía Tortugas, BA=Bahía Asunción, AB=Abreojos (see
Vogel (in prep) for a discussion of locations). The RFLP scores were generated with the
restriction enzymes AciI, DdeI, !TaqI, and AflIII. Sequence results are below the diagonal,
RFLP results are above. *Statistically significant values (p<0.05) are underlined and in
italics.
TR NR SN SC CE SQ MSD BT BA AB
TR -0.009 0.032 0.019 0.174* -0.009 -0.063 0.006 0.016 -0.011
NR 0.012 0.031 -0.037 0.156* -0.027 -0.018 -0.015 -0.013 0.037
SN 0.028 0.017 0.046 0.074 0.124 0.113 0.087 -0.029 -0.104
SC 0.077 -0.015 0.014 0.279* -0.040 0.048 -0.026 -0.076 0.025
CE -0.020 0.086* 0.053 0.153* 0.311* 0.144 0.332* 0.301* 0.128
SQ -0.010 -0.021 0.059 -0.024 0.129* 0.009 -0.123 0.004 0.093
MSD -0.007 0.020 0.109 0.032 0.074 0.038 0.086 0.104 0.080
BT -0.044 -0.054 0.009 -0.050 0.049 -0.087 0.017 0.003 0.077
BA -0.021 -0.001 -0.014 -0.050 0.102* -0.018 0.059 -0.062 -0.017
AB 0.000 0.008 -0.087 0.016 0.042 0.031 0.043 -0.028 -0.015
85
Figure 9. Plot of !
ST
scores between ten populations of Paralabrax clathratus with
sequence and RFLP data. Locations are discussed in (Vogel and Edmands, in prep-a).
86
Table 15. Results for the scoring of eight microsatellites (in boldface, first column) in the
parents and offspring of three crosses of Paralabrax clathratus. Parental genotypes are
indicated above each marker name. Crosses are identified as male $ female. Possible allele
scores, expected frequencies and observed frequencies are given for each marker in each
cross. Expected numbers were statistically significantly different for marker AV17 in cross
6$4; none were statistically different after Bonferroni correction.
Cross: 2x4 2x5 6x4
exp. obs. exp. obs. exp. obs.
genotypes freq. freq. genotypes freq. freq. genotypes freq. freq.
Microsatellite: AV15
Parental: 175/187$115/179 175/187$183/191 103/231$115/179
F1: 115/175 0.25 0.24 175/183 0.25 0.13 103/115 0.25 0.37
115/187 0.25 0.29 175/191 0.25 0.20 103/179 0.25 0.13
175/179 0.25 0.29 183/187 0.25 0.40 115/231 0.25 0.25
179/187 0.25 0.18 187/191 0.25 0.27 179/231 0.25 0.25
n: 17 15 8
significance: n.s. n.s. n.s.
Microsatellite: AV88
Parental: 166/198$158/174 166/198$162/206 162/174$158/174
F1: 158166 0.25 0.31 162166 0.25 0.47 158162 0.25 0.40
158198 0.25 0.13 162198 0.25 0.21 158174 0.25 0.10
166174 0.25 0.13 166206 0.25 0.21 162174 0.25 0.30
174198 0.25 0.44 198206 0.25 0.11 174174 0.25 0.20
n: 16 19 10
significance: n.s. n.s. n.s.
Microsatellite: AV6
Parental: 209/261$175/197 209/261$195/225 217/221$175/197
F1: 175209 0.25 0.29 195209 0.25 0.17 175217 0.25 0.25
175261 0.25 0.21 195261 0.25 0.39 175221 0.25 0.13
197209 0.25 0.36 209225 0.25 0.17 197217 0.25 0.00
197261 0.25 0.14 225261 0.25 0.28 197221 0.25 0.63
n: 14 18 8
significance: n.s. n.s. n.s.
Microsatellite: Mbo066
Parental: 97/99$97/99 97/99$99/99 97/97$97/99
F1: 9797 0.25 0.26 9799 0.50 0.56 9797 0.50 0.50
9799 0.50 0.43 9999 0.50 0.44 9799 0.50 0.50
9999 0.25 0.30
n: 23 18 16
significance: n.s. n.s. n.s.
87
Table 15. Continued.
Microsatellite: Gag010
Parental: 112/114$112/114 112/114$112/112
100/112$
112/114
F1: 112112 0.25 0.35 112112 0.50 0.63 100112 0.25 0.17
112114 0.50 0.50 112114 0.50 0.37 100114 0.25 0.06
114114 0.25 0.15 112112 0.25 0.56
112114 0.25 0.22
n: 20 19 18
significance: n.s. n.s. n.s.
Microsatellite: Pam13
Parental: 78/100$82/100 78/100$84/100 78/100$82/100
F1: 7882 0.25 0.17 7884 0.25 0.05 7882 0.25 0.05
78100 0.25 0.09 78100 0.25 0.20 78100 0.25 0.23
82100 0.25 0.35 84100 0.25 0.55 82100 0.25 0.45
100100 0.25 0.39 100100 0.25 0.20 100100 0.25 0.27
n: 23 20 22
significance: n.s. n.s. n.s.
Microsatellite: AV17
Parental: 137/137$137/141 137/137$137/141 137/159$137/141
F1: 137137 0.50 0.67 137137 0.50 0.54 137137 0.25 0.30
137141 0.50 0.33 137141 0.50 0.46 137141 0.25 0.60
137159 0.25 0.05
141159 0.25 0.05
n: 24 24 20
significance: n.s. n.s. 0.04
Microsatellite: AV115
Parental: 107/111$107/123 107/111$115/131 99/103$107/123
F1: 107107 0.25 0.26 107115 0.25 0.19 99107 0.25 0.23
107111 0.25 0.22 107131 0.25 0.24 99123 0.25 0.31
107123 0.25 0.13 111115 0.25 0.38 103107 0.25 0.31
111123 0.25 0.39 111131 0.25 0.19 103123 0.25 0.15
n: 23 21 13
significance: n.s. n.s. n.s.
88
There were no significant differences between expected and observed genotype counts after
sequential Bonferroni correction (Table 15). Because of low DNA quantity, sample sizes
ranged from 8-23 across markers (Table 15). Although the contingency tests had low power
(especially for the 6 $ 4 cross), in combination with the consistent amplification of expected
allele combinations they present a strong argument for Mendelian inheritance.
There was no evidence of linkage except for a series of significant results for Pam13 and
AV17. The marker Pam13 is generally problematic as it is unreasonable that it is singularly
linked to almost all the other loci. Caution should be taken before using this locus in other
studies, such as population analyses. Other significant results, centered on AV17, but only in
the third cross, 6 $ 4. Amplification in the third cross was generally difficult; because of no
significant comparisons in the other two crosses, we do not believe this locus is linked.
Results provide confidence in the utility of these microsatellite loci (except perhaps for
Pam13). Analyses that assume Mendelian transmission such as estimation of variance
effective population size (Pollak, 1983) or geographic population structure (F
ST
) (Wright,
1931) for example, could be validly run. Combined with the mitochondrial protocol, the loci
offer a suite of tools for analyzing Paralabrax clathratus. There is also evidence that the
microsatellite primers are conservative enough to be used in other congeners such as
Paralabrax nebulifer (G. Benavides, pers. comm.), thereby extending their utility.
Acknowledgements
Support was provided through a Rose Hills Fellowship from the Wrigley Institute for
Environmental Studies (ABV) and a NSF Dissertation Enhancement Grant (KS).
89
Chapter 3 References
ANDERSON, T. (2001) Predator responses, prey refuges, and density-dependent mortality
of a marine fish. Ecology, 82, 245-257.
BAUMS, I., HUGHES, C. & HELLBERG, M. (2005) Mendelian microsatellite loci for the
Caribbean coral Acropora palmata. Marine Ecology Progress Series, 288, 115-127.
BURTON, R. (1996) Molecular tools in marine ecology. Journal of Experimental Marine
Biology and Ecology, 200, 85-101.
CARR, M. (1994) Effects of macroalgal dynamics on recruitment of a temperate reef fish.
Ecology, 75, 1320-1333.
COPE, J. (2004) Population genetics and phylogeography of the blue rockfish (Sebastes
mystinus) from Washington to California. Canadian Journal of Fisheries and
Aquatic Science, 61, 332-342.
DE MEEUS, T., HUMAIR, P., GRUNAU, C., DELAYE, C. & RENAUD, F. (2004) Non-
Mendelian transmission of alleles at microsatellite loci: an example in Ixodes
ricinus, the vector of Lyme disease. International Journal of Parisitology, 34, 943-
950.
ESCHMEYER, W. N., HERALD, E. S. & HAMMANN, H. (1983) A Field Guide to Pacific
Coast Fishes, New York, New York, Houghton Mifflin Company.
EXCOFFIER, L., LAVAL, L. & SCHNEIDER, S. (2005) Arlequin ver. 3.0: An integrated
software package for population genetics data analysis. Evolutionary Bioinformatics
Online, 1, 47-50.
GRAVES, J. & MCDOWELL, J. (2003) Stock structure of the world's istiophorid billfishes:
a genetic perspective. Marine and Freshwater Research, 54, 287-298.
LAURENT, V., VOISIN, M. & PLANES, S. (2006) Genetic clines in the Bay of Biscay
provide estimates of migration for Sardina pilchardus. Journal of Heredity, 97, 81-
88.
LOVE, M. (1996) Probably More than You Want to Know About the Fishes of the Pacific
Coast, Santa Barbara, CA, Really Big Press.
LOWE, C. & TOPPING, D. (2003) Movement patterns, home range, and habitat utilization
of adult kelp bass Paralabrax clathratus in a temperate no-take marine reserve.
Marine Ecology Progress Series, 256, 205-216.
MORIN, P., LUIKART, G., WAYNE, R. & GROUP, T. S. W. (2004) SNPs in ecology,
evolution and conservation. Trends in Ecology and Evolution, 19, 208-216.
POLLAK, E. (1983) A new method for estimating the effective population size from allele
frequency changes. Genetics, 104, 531-548.
ROCHA-OLIVARES, A., ROSENBLATTT, R. & VETTER, R. (1999) Cryptic species of
rockfish (Sebastes, Scorpaenidae) in the southern hemisphere inferred from
mitochondrial lineages. Journal of Heredity, 90, 404-411.
ROCHA-OLIVARES, A. & VETTER, R. D. (1999) Effects of oceanographic circulation on
the gene flow, genetic structure, and phyogeography of the rosethorn rockfish
(Sebastes helvomaculatus). Canadian Journal of Fisheries and Aquatic Science, 56,
803-813.
SCHUELKE, M. (2000) An economic mathod for the fluorescent labeling of PCR
fragments. Nature Biotechnology, 18, 233-234.
SELKOE, K. & TOONEN, R. (2006) Microsatellites for ecologists: a practical guide to
using and evaluating microsatellite markers. Ecology Letters, doi: 10.111/j.1461-
0248.2006.00889.x.
90
SELKOE, K., VOGEL, A. & AFTAB, B. (2005) Eight polymorphic microsatellite markers
for kelp bass, Paralabrax clathratus, amplified in three multiplex polymerase chain
reaction sets. Molecular Ecology Notes, 5, 127-129.
SMITH, K., ALBERTS, S., BAYES, M., BRUFORD, M., ALTMANN, J. & OBER, C.
(2000) Cross-species amplification, non-invasive genotyping, and non-Mendelian
inheritance of human STRPs in Savannah baboons. American Journal of
Primatology, 51, 219-227.
TAKAGI, M., CHOW, S., OKAMURA, T., SCHOLEY, V., NAKAZAWA, A.,
MARGULES, D., WEXLER, J. & TANIGUCHI, N. (2003) Mendelian inheritance
and variation of four microsatellite DNA markers in the yellowfin tuna Thunnus
albacares. Fisheries Science, 69, 1306-1308.
THOMSON, C. (2001) Human Ecosystem Dimension. IN LEET, W., DEWEES, C.,
KLINGBEIL, R. & LARSON, E. (Eds.) California's living marine resources: A
status report., California Department of Fish and Game.
VAN HEYNINGEN, V. & YEYATI, L. (2004) Mechanisms of non-Mendelian inheritance
in genetic disease. Human Molecular Genetics, 13, 225-233.
VAN'T HOF, A., ZWANN, B., SACCHERI, I., DALY, D., BOT, A. & BRAKEFIELD, P.
(2005) Characterization of 28 microsatellite loci for the butterfly Bicyclus anynana.
Molecular Ecology Notes, 5, 169-172.
VOGEL, A. & EDMANDS, S. (in prep-a) Population genetic structure in Paralabrax
clathratus and Clinocottus analis across potential barriers to gene flow.
VOGEL, A. & EDMANDS, S. (in prep-b) Temporal and spatial genetic variance in the kelp
bass, Paralabrax clathratus.
WAPLES, R. S. & ROSENBLATT, R. H. (1987) Patterns of larval drift in southern
California marine shore fishes inferred from allozyme data. Fisheries Bulletin, 85, 1-
11.
WARD, R. D. (2000) Genetics in fisheries management. Hydrobiologia, 420, 191-201.
WRIGHT, S. (1931) Evolution in Mendelian populations. Genetics, 16, 97-159.
91
Chapter 4: Correlating processes associated with ocean triads to the recruitment of the
kelp bass, Paralabrax clathratus
Abstract
Recruitment in kelp bass, Paralabrax clathratus, is incompletely understood. We tested
Bakun’s Ocean Triad Hypothesis on recruitment of this important eastern Pacific sport
fishery species by comparing recruitment indices from seven locations to principal
component loadings of AVHRR SST satellite imagery. As predicted, recruitment in P.
clathratus typically had dome-shaped correlations to upwelling and development of
retention zones. By testing the Ocean Triads Hypothesis, we have worked to define better
the generalized and specific oceanographic events important to the life cycle of this species.
Introduction
Kelp bass (Paralabrax clathratus) are abundant in eastern Pacific kelp and rocky structure
from the southern side of the Viscaíno Peninsula, Baja California Sur, MX to Point
Conception, California, USA (Eschmeyer et al., 1983). This serranid is important to the
$300 million California sport fishing economy (Thomson, 2001, Love, 1996). Although
widely studied (Carr, 1994, Love et al., 1996, Cordes and Allen, 1997, Findlay and Allen,
2002), recruitment (metamorphosis of larvae out of the plankton) remains unpredictable.
Recruitment has been correlated to frond density of giant kelp (Macrocystis pyrifera) (Carr,
1994) and tidal bores (Findlay and Allen, 2002). However, analysis of environmental
influence is in general difficult; (Myers, 1998) found that only 1 in 47 published correlations
were used in management. Usable correlations are elusive because of multiple influences
(Myers, 1998), poor understanding of larval response (MacKenzie, 2000), autocorrelation
(Myers, 1998) and behavior changes (Bakun, 1996).
92
One response to this difficulty are ‘synthesis explanations,’ which bring multiple
connections into a single hypothesis (Cole and McGlade, 1998a). Two examples are the
‘Optimal Environmental Window’ (OEW) (Cury and Roy, 1989) and the ‘Ocean Triad
Hypthesis’ (OTH) (Bakun, 1996). The OEW theory finds recruitment to have a non-linear,
dome-shaped relationship to environmental forcers. Empirical evidence has been found in
wind-induced turbulence and Ekman upwelling affecting clupeoid recruitment (Cury and
Roy, 1989, Roy et al., 1992, Cury et al., 1995). Weak winds limit recruitment by reducing
nutrient upwelling, whereas strong winds disrupt food patches (Cury and Roy, 1989).
The OTH, which can be used with the OEW, originated from observation that larvae benefit
from ‘enrichment’ (via upwelling), but not to the degree that food patches are degraded or
larvae are transported out to sea (Bakun, 1996). ‘Concentration’ and ‘retention’ are
therefore important factors that can occur though combinations of reduced upwelling and/or
nearby zones of stability. These triad zones have been described in the Mediterranean Sea
(Agostini and Bakun, 2002) as well as correlated to engraulid and clupeoid recruitment in
the Benguela upwelling system off of southwestern Africa (Cole, 1999) and in the Sea of
Cortez (Lluch-Cota et al., 1999).
The two areas with the highest densities of kelp bass, as determined by density surveys
(Love et al., 1999) and anecdotal reports from sport fishermen (also see Kelly and Kira,
1998), are also recruitment triad zones (Figure 10, see Figure 10.1.a in Bakun, 1996) for
Engraulus mordax, the Californian northern anchovy (Cury et al., 1995). The Southern
California Bight (SCB) and the waters south of the Viscaíno Peninsula (The Gulf of Ulloa,
93
Figure 10. Surface current map along the California coast and Baja Peninsula as adapted
from Durazo and Baumgarter (2002) and Hickey et al. (2003). Seasonal timing of currents is
not explicitly shown; poleward countercurrents along the near shore for example, begin in
late summer and extend though the winter. Point Conception (and north) and the tip of the
Vizcaíno Peninsula (Punta Eugenia) are upwelling centers; below them are the protected
cyclonic gyres in the Gulf of Ulloa (GU) and Southern California Bight (SCB). Both are
present during the late summer recruitment of Parlabrax clathratus but are replaced by
equatorward moving water at other times of the year. The California Current occurs
generally to the east of the SCB and GU and is consistent in its equatorward direction.
94
GU) are situated to the south of an eastern-boundary upwelling zone (Point Conception for
the SCB, Punta Eugenia for the GU, Figure 10) but are protected from strong winds and
currents. Although eddies in these two areas are cyclonic, propagating geostrophic gyres
promote seasonal retention and concentration in both locations (Bakun, 1996).
To test the OTH for Paralabrax clathratus, we developed a recruitment index from age
distributions of 2505 samples collected at six locations (five in the SCB and one in the GU,
Figure 11). We followed a methodology created by Cole and McGlade (1998b) for
comparing recruitment indices to oceanographic indices (determined from satellite imagery).
Principal Components Analysis was used to identify oceanographic characteristics that
corresponded to triad influences.
Materials and Methods
Computer Resources
All data analysis was performed in Matlab® Version 7.1 (R14SP3) with the Curve Fitting
(1.1.4) and Statistics (5.1) Toolboxes. Statistical analysis with the Alternating Conditional
Expection (ACE) algorithm utilized the CRP Toolbox available at: http://www.agnld.uni-
potsdam.de/~marwan/toolbox/.
Sample Collection
From 2001-2003, 2515 Paralabrax clathratus samples were collected between June and
August from six locations (Figure 11, Table 16). Samples were taken from fish caught with
hook and line or from carcasses of fish caught on local sport fishing boats. Sagittal otoliths
were removed from the ventral side of the skull stored in 1.7ml microcentrifuge tubes.
95
Figure 11. Sample locations of Paralabrax clathratus and areas analyzed with Pathfinder
AVHRR images. Collection locations of fish include (sample sizes in parentheses): (A)
Abreojos, BCS, Mexico (1002); (B) Coronado Islands, BCN, Mexico (308); (C) San Diego,
CA, USA (532); (D) San Clemente Island, CA, USA (316); (E) Santa Catalina Island, CA,
USA (197); (F) North Channel Islands, CA, USA (288) and (G) Santa Barbara, CA, USA
(150). Imagery was analyzed for two locations: around the Vizcaíno Peninsula and Punta
Eugenia (25-28.5N, 112-116.1W) and the Southern California Bight (31.5-34.5N, 116.6-
120.5W).
96
Table 16. Description of areas used in remote sampling of SST data from 1985 to 2005.
Letter codes for locations correspond to Figure 11. Standard error for missing pixel
percentages is given in parentheses.
Location: Viscaíno Peninsula Southern California Bight
Coordinates: 25-28.5°N,112-116.1°W 32-37°N, 117-125°W
Average percent annual: 6.0% (±0.003) annual: 8.0% (±0.003)
missing pixels May to Sept.: 8.3% (±0.005) May to Sept.: 10.9% (±0.006)
(excl. bad weeks)
Number of bad Jan. to Mar.: 3 Jan. to Mar.: 8
weeks: Apr. to Jun.: 26 Apr. to Jun.: 14
Jul. to Sep.: 1 Jul. to Sep.: 39
Oct. to Dec: 7 Oct. to Dec: 9
Sampling locations (A) Abreojos (n=1002) (B) Coronado Islands (n=308)
(26.70°N, 113.56°W) (32.43°N, 117.26°W)
(C) San Diego (n=532)
(32.75°N, 117.29°W)
(D) San Clemente Island (n=316)
(32.83°N, 118.48°W)
(E) Santa Catalina Island (n=308)
(33.45°N, 118.49°W
(F) Northern Channel Islands
(n=288)
(33.85-34.1°N, 119.30-120.46°W)
(G) Santa Barbara (n=150)
(34.42°N, 119.96°W)
97
We also analyzed visual survey data of new recruits, collected by the Kelp Forest
Monitoring Program (Davis et al., 1997) from 1985-2004, from the Northern Channel
Islands (NCI; Figure 11, Table 16) of San Miguel (two sites), Santa Maria (three sites),
Santa Cruz (five sites) and Anacapa (three sites). Most samples came from Santa Cruz
Island (182/288) and Anacapa Island (80/288). While summing data from 13 sites spread
over four islands presented unknown variance, grouping created an index that could be
interpreted as a reflection of the Santa Barbara Channel in general (236/288 samples were
from channel-side collection sites).
Otolith preparation
Otoliths were embedded in polyester sanding resin and cut on a low speed Buhler Isomet
saw with a single 3” ! 0.006” Norton diamond wheel (B.E.E. Industrial Supply,
http://www.beeind.com). Three to four transverse sections across the nucleus with 0.30mm
thicknesses were taken from each otolith. A 25X dissecting scope was used to select one to
three slices that clearly displayed the banding. Slices were attached to a microscope slide
with cytoseal mounting medium for storage and scoring. Since banding was typically
already distinct, sections were not polished.
Otolith validation
Although banding occurs annually (Love et al., 1996), rapid first year growth prevents easy
interpretation of the first winter band. Three young-of-the-year were therefore captured in
November 2003 and soaked in 600mg/L oxytetracycline for six hours in total darkness (as in
(Cermeño et al., 2003)). Individuals were housed in 15 gallon aquariums with running
98
seawater and fed frozen squid and anchovy. In April of 2004 transverse slices of their
otoliths were visualized with ultraviolet light to compare the fluorescent oxytetracycline
mark and the first winter band.
Otolith scoring
Otoliths were read with 25X transmitted light. Year counts were based upon two transects
from the distal and dorsal edges to the nucleus. Because samples were collected in summer,
the season in which kelp bass are born (Love et al., 1996, Cordes and Allen, 1997), the final
year was counted as the last winter growth zone and the incomplete summer band on the
otolith edge.
A pilot study determined that certain years had distinct banding patterns and could be used
as landmarks. This conclusion was solidified when these patterns were found to be
staggered by year-differences in collection season. We also noted that these regular
differences in banding changed the morphology of the banding and checking around the
nucleus for individuals born in the unusual year. A series of scoring rules for each particular
year-class were therefore created to account for these patterns.
Two readers independently recorded the year counts. Discrepancies were discussed with
both readers present; data points were saved if a consensus on the disputed ages could be
reached. If not, the sample’s second otolith was scored. In 10 cases this otolith was also
unscoreable and the sample removed (leaving 2505 samples).
99
Development of recruitment index
We created a recruitment index by calculating survivorship trends for “lightly fished” and
“heavily fished” populations. Histogram plots of birth year indicated that Abreojos, the
Coronado Islands and San Clemente Island had substantial numbers of older individuals
(37.3%±15.4% born before 1991) and by inference, lower mortality rates. The other three
locations (Santa Barbara, Santa Catalina Island and San Diego) were more heavily fished
(3.0%±1.5% born before 1991).
For each group, data were log
10
transformed and the best fit line calculated. The regression
equation was used to create hypothetical values from 1969 (the birth year of the oldest
individual) to 2005. These values were transformed back by their inverse log
10
, and divided
by the 2005 value such that year zero (2005) had a survivorship of 1.0000 and all other years
were exponentially decreasing fractions. Mortality was the stepwise ratios in survivorship
between congruent years.
Recruitment for each year at adult locations was estimated by back calculating a year class
through the yearly mortality steps. These recruit values were then normalized (Figure 12).
Low sample years were removed from the sample set, so that data ranged from 1985 to 1997
for the lightly fished populations and 1989 to 1997 for two of the heavily fished populations
(Santa Catalina Island and San Diego) and 1991 to 1997 for the third (Santa Barbara). No
mortality was removed before normalization of the 1985-2004 NCI visual survey data
(Figure 13).
100
Figure 12. Plot of Recruitment Indices for the six locations sampled in this study.
101
Figure 13. Plot of Recruitment Index for the Northern Channel Islands from data collected
as part of the Kelp Forest Monitoring Program.
102
Oceanographic data
We used Version 5.0 of the NOAA Advanced Very-High-Resolution Radiometer (AVHRR)
Oceans Pathfinder Sea Surface Temperature (SST) data set from the Nation Oceans Data
Center and the Rosenstiel School for Marine and Atmospheric Science. These data are
distributed by the Physical Oceanography Distributed Active Archive Center, have 4km
resolution and range from 1985 to present.
Pixels for the GU and the SCB were extracted from eight-day-average SST images for
January 8
th
, 1985 to December 26
th
, 2005 (Figure 11, Table 16). Cloud and land pixels were
marked through quality flag overlay (range 1:7; <4 were marked as missing), after which
ascending and descending images were combined. If a pixel had two good values (one from
each image), the values were averaged. Plots of temperature ranges found two inaccurate
cloud masks; these images were removed.
This method contradicts work that used only descending pass data to avoid daytime heating
of the thin surface layer (Nezlin and McWilliams, 2003). Because seasonal variance was
described by Principal Components Analysis (see below) though, the analysis benefited from
improved coverage and was robust at finding seasonal trends in the first few components
while relegating noise variance to later ones.
Each combined image was cleaned of missing pixels. If at least 40% of ocean pixels were
missing, the image was marked. Otherwise, a repeating algorithm determined the values of
the eight pixels surrounding each missing pixel (a single-pixel border of land values
permitted analysis of edge pixels). If at least three of these pixels contained good data, the
103
missing pixel assumed the average value of the good quality, neighboring pixels. The
average percentage of missing pixels for both annual and summer data are presented in
Table 16.
For both extracted sets, images were vectorized and concatenated such that the vertical axis
was geographical and the horizontal axis represented time in 945 vertical vectors. If a
vertical vector with marked, bad data was found to both follow and proceed good weeks, it
was recalculated as the average of the two good vectors. Longer periods of bad data were
removed (Table 16) including images for much of July and August 1995; that entire year
was removed from analysis. Land, identified by horizontal vectors with less than 60% of its
pixels with high quality values, was also eliminated. Finally, pixels for the Sea of Cortez,
which were found in images of Baja, Mexico and the GU, were excluded.
Principal Components Analysis (PCA) and development of oceanographic index
Methodology generally followed Cole and McGlade (1998b). To avoid bias from high
variability images (Cole and McGlade, 1998b), each eight-day image was normalized.
Principal Components Analysis (PCA) was performed such that Eigenvector loadings
described temporal variability. We analyzed the second and third components, viewed as
most likely to be rule obeying (Sirabella et al., 2001). Component one was invariably
positive and represented average structure (as noted by Cole, 1999). Components four and
five were also initially surveyed, but were reflective of noise. To define the connections of
components to oceanographic patterns, we took images that corresponded to the 15 highest
loadings, the 15 lowest, and the 15 closest to zero. These were searched for groups that had
repetitive patterns from summer (recruitment season) months.
104
Because components two and three for GU imagery connected to different oceanographic
patterns, two GU oceanographic indices were created from summing the loadings of each
component independently. The two components from imagery of the SCB described similar
processes found in distinct areas. The Santa Barbara and NCI locations was therefore
compared to a single index created from both components and the other SCB populations
compared only to the sum of loadings from component three. The oceanographic index for
the Santa Barbara and NCI locations included summed values for component three and
absolute values of negative loadings for component two. This method created an index that
described upwelling/retention around the Santa Barbara Channel (described below).
For all components, indices were created from loadings that occurred between May and
September. This time range corresponded to active feeding/spawning season of Paralabrax
clathratus that begins in late spring/early summer with the warming of the water and ends
with arrival of recruits in August/September.
Data Transformations and Correlation Analyses
Indices were transformed with an Alternating Conditional Expectation (ACE) algorithm,
which transforms a dependent variable and n independent variables to maximize correlation
(Breiman and Friedman, 1985). This method has been used successfully to compare
recruitment to PCA loadings (Cole, 1999) and to visualize the dome-shaped OEW (Cury and
Roy, 1989, Cury et al., 1995). Transformation requires no a priori assumptions about data
relationships, and so is useful for analyzing data that are potentially dome-shaped. Statistical
significance was determined with a linear regression of the transformed dependent and
105
independent variables, the relationship of which was visualized with a plot of the
untransformed versus transformed oceanographic indices. Report of R
2
results indicates the
amount of variance in the recruitment index that is explained by the plotted relationship.
Results
Recruitment Indices
Age structure for the 2001 and 2002/3 sampling seasons was correlated (Table 17a) and
justified grouping otolith samples from multiple collection years. Table 17b presents
correlations of raw data from the six adult samples. These results confirm the patterns in
Figure 12 (letters correspond to locations in Figure 11): age structure for (B) San Diego (C),
Santa Catalina Island (E), and to some extent Santa Barbara (G) is similar, whereas it is
unique for Abreojos (A) and San Clemente Island (D).
The NCI, with strong recruitment in 1985, 1991, 1993 and 1995, showed relationship to the
group of four similar locations (Figure 13). None of the indices matched larval counts for
Paralabrax spp. presented by Moser et al. (2001). Their data cannot be used to understand
recruitment strength because of inclusion of larvae from the sister species, Paralabrax
nebulifer and P. maculatofasciatus (Moser et al., 2001).
There could be autocorrelation the data because of single recruitment events potentially
feeding multiple locations. At a minimum, differences between the three recruitment
patterns (Abreojos, San Clemente Island, all other samples) can be judged. Every site
though, had unique characteristics. Santa Barbara for example, had a strong 1994 year class
and the Coronado Islands had a strong 1990 year class.
106
Table 17a. Correlations between age class distributions for samples collected in 2001 versus
2002/3. Letters represent locations as correspond to Figure 11. Because (1) sampling in
2002/3 could collect an extra year or two of younger cohorts than in 2001 and (2) sample
size affected the distribution of rare, older age classes, correlations were limited to the age
classes effectively sampled by both years. *The correlation between San Clemente Island
samples (D) for example, increases to R
2
=0.6678 (p = 0.0039) when comparison only
includes the main part of the sample from 1988-1997.
year class
Location range compared R
2
p value
(A) Abreojos 1976-1997 0.9189 <0.0001
(B) Coronado Islands 1987-1997 0.8341 <0.0001
(D) San Clemente Is. 1982*-1997 0.2536 0.0467
(E) Santa Catalina Is. 1992-1996 0.8492 0.0261
All Samples 1975-1997 0.6700 <0.0001
107
Table 17b. Pairwise R
2
values for linear correlations of raw data from the 6 adult sampling
locations. Letters represent locations as correspond to Figure 11. Results indicate similarity
between the Coronado Islands (B), San Diego (C), Santa Catalina Island (E) and Santa
Barbara (G). Abreojos (A) and San Clemente Island (D) appear unique. ***p ! 0.001, **p
! 0.01.
A B C D E G
A - 0.1134 0.0776 0.0951 0.1308 0.0245
B - 0.7772*** 0.0546 0.8639*** 0.2713**
C - 0.1345 0.7275*** 0.5138**
D - 0.1063 0.0089
E - 0.2286**
G -
108
Imagery
Plots of component loadings with seven point moving averages are presented in Figure 14.
Both components for the GU have a seasonal signal; spectral density analysis with a Fourier
transformation found single peaks at one year (Figure 15). Only component three for the
SCB had an annual signal. Component two had two minima per year, approximately in late
spring and early winter (Figure 14).
Component two from GU images associated with warm southern water (Figures 16-17, see
Figure 10). The component was positively correlated to El Niño Southern Oscillation
(ENSO) events (R
2
=0.323, p-value=0.0089; ENSO index was a sum of monthly values from
the nino3 index [International Research Institute: http://iri.columbia.edu/cem/]). The
variability tracked by this component and component three (described below) are equivalent
to the orthogonal functions of (Espinosa-Carreon et al., 2004) but are switched, because of
different image areas.
Imagery of the entire Baja Peninsula (not shown) demonstrated the warming to involve both
coastal heating (as noted by (Espinosa-Carreon et al., 2004)) and poleward movement of
water that originated south of the Baja Peninsula. The warm water propagated up the coast
with much of it moving westward and into the path of the California Current with contact of
Bahia Magdalena (Figure 10). Some water continued northward until it hit the GU. This
poleward water, depending on the ENSO state, could be driven by either Tropical Surface
Water (no ENSO) or Subtropical Surface Water (ENSO) (Durazo and Baumgartner, 2002).
109
Figure 14. Plots of loading values for principal components 2 and 3 from images of the
Viscaíno Peninsula and Southern California Bight (SCB). The year 1995 has been removed
because a large number of bad images. The line represents a 7 point moving average of the
data.
110
Figure 15. Spectral density plots of principal components 2 and 3 from images of the
Viscaíno Peninsula and Southern California Bight (SCB). Periodicity was found to be
seasonal (1 year) for all components except component 2 of the SCB.
111
Figure 16. Representative imagery of the Gulf of Ulloa that corresponded to maximal values
of principal components 2 and 3. Positive values of component 2 were interpreted as
retention of a warm water zone between the point and the coast. Positive values of
component 3 were was interpreted as Ekman upwelling. The arrow points to Punta Prieta,
an apparent southern limit to upwelling. The northern limit of upwelling is the tip of the
Viscaíno Peninsula, Punta Eugenia.
112
Figure 17. Derived principal component images for the Gulf of Ulloa. Components are
organized by row, with each image including average loadings from May 1
st
to September
30
th
. The three years chosen represent the recruitment success dome in Figure 11; 1994 is
found on the left side of the dome because of cooler than average water along the coast
(component two) and high levels of upwelling (component three). The year 1991 occurs on
the right side of the dome because of excessive warming and high upwelling. The year 1989
represents a dominant recruitment year, with moderate arrival of warm southern water and
low levels of upwelling. A combination of components two and three (14.7% of variance)
shows a mean seasonal signal, but loses the staggered timing of the components.
113
One difference with the work of Durazo and Baumgartner (2002) is that we note the
initiation of water movement in July, as opposed to fall/winter. Because this northward
movement begins south of the Baja Peninsula, we are also in contradiction with Legaard and
Thomas (2006). They interpret SST variability around the GU to indicate that warm water
enters from the west, opposite of what we found.
This pattern is similar to the Southern California Counter Current (SCCC), which moves
pole-ward along the coast of the SCB in summer months. Although the SCCC is influenced
by an eddy caused by the California Current, both are hindered by prominent land masses
connected to upwelling (in the case of the SCCC, Point Conception). The SCCC strengthens
with relaxation of local wind stress (Hickey et al., 2003); relaxation of wind, as noted for
component three below, occurs synchronously with arrival of the warm southern water.
Component three tracked Eckman upwelling (Figures 16-17) in water adjacent to the coast
and located between the tip of the Vizcaíno Peninsula and Punta Prieta (arrow, Figure 16)
(noted by Bakun and Nelson, 1977). Like component three loading values (Figure 14), wind
speeds (Pacific Fisheries Environmental Laboratory [PFEL] data;
http://www.pfeg.noaa.gov/javamenu.html) increased through the spring and peaked in June
with an average of 7.0 m/s. Wind speeds then decreased through the rest of the year.
Importantly, wind direction also changed. Assuming north to be 0° and changing angle in a
clockwise direction, the peninsula’s southern coastline is about 135°. The average wind
angles through June were between 136 and 138°, almost parallel to the coast. In July an
abrupt transition would occur where wind became increasingly southward blowing at angles
that averaged between 150 and 160°.
114
Upwelled water can reach the tip of the Baja Peninsula (imagery not shown). The extent of
the plume is reminiscent of upwelling at Point Conception and north; both locations have
winds parallel to the coast that encounter multiple points (noted for the California coast by
Miller et al., 1999). Between Punta Eugenia and P. Prieta, no less than six points (P.
Rompiente, Cabo Thurloe, Morro Hermoso, P. San Pablo, P. San Roque and P. Asunción)
appear capable of causing upwelling.
During the recruitment season for kelp bass, component three upwelling trapped the
component two, warm water against the coast in a gyre (Figure 10, seen in Figure 16). This
zone has the triad geography posited by (Bakun, 1996): (1) upwelling to the west/north, (2)
contained warm water area, (3) coast. It allows for enrichment from upwelling,
concentration with convergence of cold, upwelled water and warm, poleward-moving water,
and retention from containment of the warm water mass against the coast. In general, Gulf
of Ulloa is an export area (Del Monte Luna, 2004). Only for a short window of time in the
late summer, peak recruitment season of Paralabrax clathratus is a retention zone created by
poleward water movement. In any other part of the year mean water flow is towards the
equator.
Components also tracked upwelling in the SCB. Although neither component correlated to
ENSO events, as expected given the variable influence of El Niño on the California coast
region (Mendessohn et al., 2003), El Niño could influence the indices. The lowest three
values (less cold water) of the index occurred in 1986, 1992 and 1993 and the highest three
(increased upwelling) in 1985, 1990 and 1997. ENSO events terminated the upwelling
115
signal in these three lowest index values. End-of-summer relaxation values of component
three usually appeared in October, but by June 17
th
, 1993, component three had relaxation
loadings; for 1992 this occurred by July 18
th
and for 1986 by August 28
th
. Other El Niño
years like 1987, 1991 and 1997 still recorded high values for the component three index.
Component two behaved like the orthogonal function described by (Nezlin and McWilliams,
2003). Negative values of this component connected to cold water at Point Conception and
in the Santa Barbara Channel (Figure 18, white arrow point to Point Conception; Figure 19).
These oceanographic patterns were not seasonally confined, although most upwelling
tracked by this component appeared to occur in late spring and early winter (Figure 14). As
summer progressed, upwelling was met by warm water entering the Santa Barbara Channel
from the east. These oceanographic patterns were equivalent to those described in the GU,
with enrichment, concentration/convergence, and retention within the channel.
Component three for the SCB indicated development of a cool, north/south column from
seasonal upwelling at Point Conception and north to Monterey (Figures 18-19). Cold water
would upwell near the coast, become part of the California Current and move south. As this
southward jet drifts west, eddies and the warm-water SCCC appear (Di Lorenzo, 2003).
Like Vizcaíno Peninsula upwelling, cold water trapped warmer water into a summer
retention zone in the SCB. Maximal loadings occurred in August/September at which point
the SCCC is moving warm water up the coast and into the Santa Barbara Channel (Harms
and Winant, 1998, Di Lorenzo, 2003).
116
Figure 18. Representative imagery of the Southern California Bight (SCB) that corresponded
to minimal values of component 2 and maximal values of principal components 3. Negative
values of component 2 represented upwelling. Positive values of component 3 described
seasonal Ekman upwelling. The arrow indicates Point Conception, the upwelling boundary
between the SCB and the central Californian coast.
117
Figure 19. Derived principal component images for the Southern California Bight (SCB).
Components are organized by row, with each image including average loadings from May 1
st
to September 30
th
. The three years chosen represent the recruitment success domes in Figure
21; part (a) for the four locations in the SCB that were compared only to component two and
part (b) for the two locations in the Santa Barbara Channel that were compared to an index
of both components 2 and 3. The years 1992 and 2001 (part a and b, respectively), from the
left side of the domes, had very little upwelling. High levels of upwelling were seen in 1997
and 2004, whereas moderate levels were found in the good recruitment years of 1989 and
1985. For part b, the high upwelling year in 1997 also represented high recruitment at the
San Clemente Island sample, in contradiction to the other locations.
118
Correlations between recruitment and principal components of GU imagery
Abreojos recruitment correlated to both oceanographic indices (Table 18). The warm water
signal (component two) was dome-shaped (Figure 20), with median values most conducive
for recruitment (Figure 17). Low index values corresponded to seasons of cool water. High
index values represented extensive warming of the GU but with reduced nutrient input and
containment by upwelling. Median values demonstrated enough heating to ‘push away’
cooler water and create a retention zone which could benefit from upwelling enrichment.
Timing (described below) was also important (it is averaged out of the images in Figure 17).
Component three had a negative pattern (Figure 20), indicating increased upwelling broke up
retention created by the component two, warm water intrusion and hurt recruitment. As
determined by imagery (and component three), every year had an upwelling signal but did
not have a clear retention pattern. A number of years, such as 1994, showed very little or
unusually timed warming in the GU (Figure 17). Other years like 1991 had overwhelming
warming, which spread early and quickly out of the gulf and towards Punta Eugenia.
The timing of heating was also important to recruitment success. Each individual month of
the component two index (May to September) showed dome-shaped relationships to the
recruitment index. Too much heat in May or too much cold in August had the same
detrimental effects of not creating a July/August retention zone.
Correlations between recruitment and principal components of SCB imagery
All sampling locations in the SCB except San Clemente Island had a dome relationship with
the principal component index (Table 18, Figure 21). The median nature of
119
Table 18. Results from correlations between recruitment indices and indices of the second
and third principal components. Imagery indices were created from yearly summations of
loadings that corresponded to the months of the summer season (May to September).
Putative description of each component is given before results. ‘Recruitment Data’ is the
location from which samples were taken to create a particular recruitment index. ‘Years’ are
the years analyzed and were limited by the recruitment indices (1995 was removed because
of poor satellite imagery). The R
2
and p-values refer to the linear correlations between the
transformed indices for the principal component and recruitment. ‘Shape of Relationship’
references the scatter plot shape of the dependent variable (Eigenvector loadings index)
versus its transformed values which can be seen in Figures 20-21.
Punta Eugenia/Viscaíno Peninsula
Component 2 (C2): Seasonal; peaks in Sep/Oct; warm water against coast; 10.5% of
variance
Component 3 (C3): Seasonal; peaks in Jul/Aug; upwelling on southern side of the
peninsula, but north of Abreojos; 4.2% of variance
Shape of
Recruitment Data Years R
2
p-value Relationship
Abreojos 1985-1997 0.844 <0.0001 C2: dome
0.912 <0.0001 C3: negative
Southern California Bight
Component 2 (C2): Undefined cycle length; minima correspond to upwelling, maxima
to relaxation and warming of upwelling zones; 5.3% of variance
Component 3 (C3): Seasonal; peaks in May/Jun; seasonal upwelling cycle with
retention of warm water in bight; 3.9% of variance
Shape of
Recruitment Data Years R
2
p-value Relationship
Coronado Islands 1985-1997 0.693 0.0008 C3: dome
San Diego 1989-1997 0.455 0.0660 C3: dome
San Clemente Is. 1985-1997 0.745 0.0003 C3: positive
Santa Catalina Is. 1989-1997 0.760 0.0048 C3: dome
Santa Barbara 1991-1997 0.962 <0.0001 C2/C3: dome
N. Channel Is. 1985-2004 0.275 0.021 C2/C3: dome
120
Figure 20. Scatter plots of the Oceanographic Index from imagery of the Gulf of Ulloa
versus their ACE transformed values for comparisons with the Abreojos recruitment
sampling location (‘A’ in Figure 11).
121
Figure 21. Scatter plots of the Oceanographic Index from imagery of the Southern California
Bight versus their ACE transformed values. Letters indicate locations as corresponding to
Figure 11: (B) Coronado Islands; (C) San Diego; (D) San Clemente Island; (E) Santa
Catalina Island; (F) Northern Channel Islands; (G) Santa Barbara. Locations B-E are were
compared to principal component 3; locations F and G were a sum of component 3 and
upwelling loadings from component 2 (loadings were transformed to their sign-inverse
values).
122
the successful recruitment years can be seen in the derived imagery of Figure 19. Perhaps
other influences such as ENSO events, when timed appropriately, moderate the upwelling
followed by component three at San Clemente Island. Alternatively, this pattern might
indicate exportation differences between locations, a topic to which we shall return in the
Discussion.
Because of their position in the Santa Barbara Channel, the Santa Barbara and NCI locations
undergo hydrological influences unique to the SCB (see, for example, Harms and Winant,
1998). For that reason we included the loadings of component two into these locations’
oceanographic indices. By changing the sign of component two, increase in the index point
value represented an increase in cool water.
Like 3/4 of the other SCB locations, Santa Barbara and NCI recruitment had dome
relationships with the oceanographic index (Figure 21). Both components indicated a west
to east, cool to warm-water gradient as upwelled water entered the Santa Barbara Channel
from the western/Pacific side and the SCCC entered from the east (Figure 18). Upwelled
water moves along the southern boundary, while the SCCC passes along the mainland coast
to the north (Harms and Winant, 1998). The shear from these opposite currents is greatest in
summer and can create cyclonic eddies within the channel (Harms and Winant, 1998).
Median levels of these events could promote retention and concentration of Paralabrax
clathratus larvae with planktonic food sources, as well as with frontal zones (which attract
larval and juvenile fish [Bakun, 1996, MacKenzie, 2000]). Although P. clathratus larvae
were not described, this gyre concentrates other species (Nishimoto and Washburn, 2002).
123
Discussion
Recruitment Correlations
Although correlations do not prove process, we have found a positive result testing a theory
developed with engraulids and clupeoids, fish with very different lifecycles than Paralabrax
clathratus. Studying the relevance of a recruitment mechanism to multiple species improves
understanding of difficult-to-define generalities in recruitment (Myers, 1998). Previously,
only the Californian northern anchovy (Engraulis mordax) had been tested in these locations
for correlation with ocean triads (Cury et al., 1995). Like the analysis performed with E.
mordax, recruitment had a nonlinear, dome-shaped relationship to upwelling/retention.
The SCB and the GU have a triad of processes that promote larval survival (Bakun, 1996).
Both areas have north/northwest sources of upwelled nutrients and zones to the
south/southeast that promote concentration and retention. The gyres in these locations (and
the Santa Barbara Channel) are cyclonic and imply a dispersive environment, in
contradiction to the explanations presented here. However, horizontal flows are small and
can be overcome by zooplankton in search of food in upwelling-inducing, cyclonic gyres
(Nishimoto and Washburn, 2002). Propagating geostrophic gyres can also promote retention
in calm water areas south of upwelling zones in eastern boundary currents (Bakun, 1996).
Recruitment in the GU correlated to a retention process where upwelling, the peninsula and
a poleward current created a convergence zone and trapped warm water. Too little current,
and convergence was minimal, too much current and the upwelling zone was overrun. With
a well-timed, moderate arrival of warm water, a well-formed zone benefited from
enrichment (upwelling), concentration (convergence) and retention (barrier to current flow).
124
In the SCB, recruitment also had, except for the San Clemente Island sample, a dome-shaped
relationship to upwelling and retention. Cold water, from upwelling at Point Conception and
to the north, moved equatorward and isolated a warm coastal zone. The warm water, like
that of the GU, was poleward moving. While these processes promoted enrichment and
retention, we have not explicitly described a concentration pattern within the main body of
the SCB. It is only in the Santa Barbara Channel where there was evidence of cold and
warm water convergence. Because of the complex structure of the SCB, convergence zones
might be smaller and locally defined, as discussed below.
Our recruitment index for the NCI behaved as predicted. It had high levels of variance (as
seen by the low R
2
value) that can be ascribed to inclusion of multiple locations. It too
though, showed a dome relationship between recruitment and upwelling processes.
Although not perfectly equal (the Santa Barbara index for example, has its own unique
recruitment pattern), this result provides evidence that there are generalized patterns in the
Santa Barbara Channel that broadly affect recruitment of kelp bass.
Chlorophyll Maxima
If the retention zones presented here are important areas of enrichment, concentration and
retention, they should exhibit evidence of increased photosynthesis. For the GU, this
prediction is supported by data from (Espinosa-Carreon et al., 2004). An orthogonal
function analysis found a mode that tracked increased concentration of chlorophyll in the
convergence/retention zone.
125
The SCB is more complicated with less surface chlorophyll in the bight than in the upwelled
water to the west (Eppley, 1992). In the Santa Barbara Channel, chlorophyll concentrations
are maximal in (1) the convergence of cold water from Point Conception upwelling and the
warm water of the SCCC and (2) the gyre created by the two currents (Caldeira et al., 2005).
There is also increased chlorophyll in the shallow shelf zone along the coast (Eppley, 1992,
Caldeira et al., 2005). For deeper parts of the SCB with less surface production, isolated
areas of increased chlorophyll are associated with island wakes (Caldeira et al., 2005).
These locations are good candidates for Paralabrax clathratus larval concentration zones.
Two lines of evidence reject sub-surface maxima as concentration zones. First, artificial
collectors primarily attract recruits of water column fish along the California coast near the
surface (Ammann, 2004), a pattern found in Paralabrax clathratus (Steele et al., 2002).
Second, Moser et al. (2001) report high catches of larval Paralabrax spp. at stations which
correspond to the above mentioned near-surface zones (see their Figure 20). In the Santa
Barbara Channel, the convergence area had high catch rates, as opposed to those to the west
and directly involved with Point Conception upwelling. Their only deep water (>1000m)
location to show evidence of high larval concentration was the Santa Barbara Channel
station that corresponded to the gyre created by upwelling and SCCC, noted by Nishimoto
and Washburn (2002) to retain juveniles.
Export vs. Retention
While there is a relationship between recruitment and the OTH, the causes of variability
between the seven sites are unclear. Chief among the candidates is the tension between
export of larvae and localized recruitment, a topic that has received attention through
126
analyses of otolith microchemistry (Newman et al., 2000, Forrester and Swearer, 2002,
Gillanders, 2002). Although these recruitment patterns would be complicated by adult
movement, our multiple-year sample demonstrated stable age distributions with no evidence
of substantial adult migration. Tagging with radio transmitters has also shown home range
affinities in adults (Lowe and Topping, 2003).
The export/retention question can be seen through the mix of patterns in our recruitment
index. On the one hand, multiple recruitment indices were equivalently related to single
oceanic events. Certain year-class peaks could also be seen in different recruitment indices.
The years 1991, 1993 and 1995 for example, were good recruitment years for SCB sample
locations. On the other hand, no two recruitment indices were exactly the same; San
Clemente Island and Abreojos had especially unique age structures.
From this framework, potential concentration/retention zones could be judged by their
coverage and potential sources and destinations of recruits. San Clemente Island (and
potentially the Coronado Islands) should have a more developed retention zone (such as an
island wake) in 1996, 1991, 1985, 1990 and 1997, when its recruitment had an increasingly
positive relationship to our oceanographic index and our other SCB locations had negative
relationships. A corollary hypothesis is that the location of the GU convergence zone should
correlate to peak recruitment along the southern coast of the Vizcaíno Peninsula.
Implications for management
One potential management tool from this research is that strong years of recruitment can be
identified with derived imagery. In the GU, the year 1989 has, over our index time period of
127
May to September, median levels of warming by southern waters (component two) and low
levels of upwelling (component three) (Figure 17). The SCB is more complicated, but there,
too, can annual differences be clearly seen with components two and three. Upwelling
around Point Conception (component two) and along the California coast (component three)
can make for strong patterns of retention (1989, 1985) or lack thereof (1992, 2001, 2004)
(Figure 19). The year 1997 presents yet another type of image, but one that had variable
results. Understanding of localized processes might help explain this variability.
It is daunting to realize though, that even broader scale cycles may be important.
MacKenzie (2000) proposed that conflicting positive and negative correlations may be
caused by limited time series data viewed from opposite sides of a dome-shaped curve.
Regime shifts, such as for Engraulids and Clupeoids, can also reverse decade-long
correlations (Bakun, 1996). The 1997/1998 El niño has been implicated as such a regime
shift (Peterson and Schwing, 2003), and it remains an open topic on how this event might
change or reinforce recruitment patterns determined in warmer decades.
Acknowledgements
Otolith preparation was facilitated by O. Nishisaki-Sosa and J. Rosales-Cassein, the Cailliet
lab and D. Busatto. The Abreojos Fishing Cooperative, H&M Landing in San Diego, the
Pierpoint and 22
nd
Street Landings in Los Angeles and J. Diamond and the Stardust out of
Santa Barbara helped with sample collection. Recruit data from the Northern Channel
Islands were provided by D. Kushner and the National Park Service. A Boren Fellowship
from the National Security Education Program provided financial support for this research.
128
Chapter 4 References
AGOSTINI, V. & BAKUN, A. (2002) 'Ocean Triads' in the Mediterranean Sea: physical
mechanisms potentially structuring reproductive habitat suitability (with example
application to European anchovy, Engraulis encrasicolus. Fisheries Oceanography,
11, 129-142.
AMMANN, A. (2004) SMURFs: standard monitoring units for the recruitment of temperate
reef fishes. Journal of Experimental Biology and Ecology, 299, 135-154.
BAKUN, A. (1996) Patterns in the ocean: Ocean processes and marine population
dynamics, La Paz, BCS, Mexico, California Sea Grant/CIB.
BAKUN, A. & NELSON, C. (1977) Climatology of upwelling related processes off Baja
California. California Cooperative Fisheries Investigations, Reports, 19, 107-127.
BREIMAN, L. & FRIEDMAN, J. (1985) Estimating optimal transformations for multiple
regression and correlation. Journal of the American Statistical Association, 80, 580-
598.
CALDEIRA, R., MARCHESIELLO, P., NEZLIN, N., DIGIACOMO, P. & MCWILLIAMS,
J. (2005) Island wakes in the Southern California Bight. Journal of Geophysical
Research-Oceans, 110, doi:10.1029/2004JC002675.
CARR, M. (1994) Effects of macroalgal dynamics on recruitment of a temperate reef fish.
Ecology, 75, 1320-1333.
CERMEÑO, P., URIARTE, A., DE MURGUÍA, A. & MORALES-NIN, B. (2003)
Validation of daily increment formation in otoliths of juvenile and adult European
anchovy. Journal of Fish Biology, 62, 679-691.
COLE, J. (1999) Environmental conditions, satellite imagery, and clupeoid recruitment in
the northern Benguela upwelling system. Fisheries Oceanography, 8, 25-38.
COLE, J. & MCGLADE, J. (1998a) Clupeoid population variability, the environment and
satellite imagery in coastal upwelling systems. Reviews in Fish Biology and
Fisheries, 8, 445-471.
COLE, J. & MCGLADE, J. (1998b) Temporal and spatial patterning of sea surface
temperature in the northern Benguela upwelling system: possible environmental
indicators of clupeoid production. South African Journal of Marine Science, 19,
143-157.
CORDES, J. F. & ALLEN, L. G. (1997) Estimates of age, growth, and settlement from
otoliths of young-of-the-year kelp bass (Paralabrax clathratus). Bulletin of the
Southern California Academy of Sciences, 96, 43-60.
CURY, P. & ROY, C. (1989) Optimal environmental window and pelagic fish recruitment
success in upwell areas. Canadian Journal of Fisheries and Aquatic Science, 46,
670-680.
CURY, P., ROY, C., MENDESSOHN, R., BAKUN, A., HUSBY, D. & PARRISH, R.
(1995) Moderate is better: exploring nonlinear climatic effects on the Californian
northern anchovy (Engraulis mordax). Canadian Journal of Fisheries and Aquatic
Science, 121, 417-424.
DAVIS, G., KUSHNER, D., MONDRAGON, J., MONDRAGON, J., LERMA, D. &
RICHARDS, D. (1997) Kelp forest monitoring handbook volume 1: sampling
protocol. Ventura, California, National Park Service, Channel Islands National Park.
DEL MONTE LUNA, P. (2004) Caracterización del centro de actividad biológica del Golfo
de Ulloa, B.C.S., bajo un enfoque del modelación ecológica. Instituto Politécnico
Nacional. La Paz, B.C.S., CICIMAR.
129
DI LORENZO, E. (2003) Seasonal dynamics of the surface circulation in the Southern
California Current system. Deep-Sea Research II, 50, 2371-2388.
DURAZO, R. & BAUMGARTNER, T. (2002) Evolution of oceanographic conditions off
Baja California: 1997-1999. Progress in Oceanography, 54, 7-31.
EPPLEY, R. (1992) Photosynthesis and new production in the Southern California Bight.
Progress in Oceanography, 30, 117-150.
ESCHMEYER, W. N., HERALD, E. S. & HAMMANN, H. (1983) A Field Guide to Pacific
Coast Fishes, New York, New York, Houghton Mifflin Company.
ESPINOSA-CARREON, T., STRUB, P., BEIER, E., OCAMPO-TORRES, F. &
GAXIOLA-CASTRO, G. (2004) Seasonal and interannual variability of satellite-
derived chlorophyll piment, surface height, and temperature off Baja California.
Journal of Geophysical Research-Oceans, 109, C03039,
doi:10.1029/2003JC002105.
FINDLAY, A. & ALLEN, L. G. (2002) Temporal patterns of settlement in the temperate
reef fish Paralabrax clathratus. Marine Ecology Progress Series, 238, 237-248.
FORRESTER, G. & SWEARER, S. E. (2002) Trace elements in otoliths indicate the use of
open-coast versus bay nursery habitats by juvenile California halibut. Marine
Ecology Progress Series, 241, 201-213.
GILLANDERS, B. (2002) Temporal and spatial variability in elemental composition of
otoliths: implications for determining stock identity and connectivity of populations.
Canadian Journal of Fisheries and Aquatic Science, 59, 669-679.
HARMS, S. & WINANT, C. (1998) Characteristic patterns of the circulation in the Santa
Barbara Channel. Journal of Geophysical Research-Oceans, 103, 3041-3065.
HICKEY, B., DOBBINS, E. & ALLEN, S. (2003) Local and remote forcing of currents and
temperature in the central Southern California Bight. Journal of Geophysical
Research-Oceans, 108, doi:10.1029/200JC000313.
KELLY, N. & KIRA, G. (1998) The Baja catch: A fishing, travel and remote camping
manual for Baja California, Valley Center, CA, Apples and Oranges, Inc.
LEGAARD, K. & THOMAS, A. (2006) Spatial patterns in seasonal and interannual
variability of chlorophyll and sea surface temperature in the California Current.
Journal of Geophysical Research-Oceans, 111, doi:10.1029/2005JC003282.
LLUCH-COTA, S., LLUCH-COTA, D., LLUCH-BELDA, D., NEVAREZ-MARTINEZ,
M., PARES-SIERRA, A. & HERNANDEZ-VAZQUEZ, S. (1999) Variability of
sardine catch as related to enrichment, concentration and retention processes in the
central Gulf of California. California Cooperative Fisheries Investigations, Reports,
40, 184-190.
LOVE, M. (1996) Probably More than You Want to Know About the Fishes of the Pacific
Coast, Santa Barbara, CA, Really Big Press.
LOVE, M., NISHIMOTO, M., SCHROEDER, D. & CASELLE, J. (1999) The ecological
role of natural reefs and oil and gas production platforms on rocky reef fishes in
southern California: Final Interim Report., U.S. Department of the Interior, U.S.
Geological Survey, Biological Resources Division.
LOVE, M. S., BROOKS, A., BUSATTO, D., STEPHENS, J. & GREGORY, P. A. (1996)
Aspects of the life histories of the kelp bass, Paralabrax clathratus, and the barred
sand bass, P. nebulifer, from the southern California Bight. Fishery Bulletin, 94,
472-481.
130
LOWE, C. & TOPPING, D. (2003) Movement patterns, home range, and habitat utilization
of adult kelp bass Paralabrax clathratus in a temperate no-take marine reserve.
Marine Ecology Progress Series, 256, 205-216.
MACKENZIE, B. (2000) Turbulence, larval fish ecology and fisheries recruitment: a review
of field studies. Oceanologica Acta, 23, 357-375.
MENDESSOHN, R., SCHWING, F. & BOGRAD, S. (2003) Spatial structure of subsurface
temperature variability in the California Current, 1950-1993. Journal of Geophysical
Research-Oceans, 108, 3093, doi:10.1029/2002JC001568.
MILLER, A., MCWILLIAMS, J., SCHNEIDER, N., ALLEN, J., BARTH, J.,
BEARDSLEY, R., CHAVEZ, F., CHERESKIN, T., EDWARDS, C., HANEY, R.,
KELLY, K., KINDLE, J., LY, L., MOISAN, J., NOBLE, M., NIILER, P., OEY, L.,
SCHWING, F., SHEARMAN, R. & SWENSON, M. (1999) Observing and
modeling the California Current System. EOS, Transactions, American Geophysical
Union, 80, 533-539.
MOSER, H., CHARTER, R., WATSON, W., AMBROSE, D., HILL, K., SMITH, P.,
BUTLER, E., SANDKNOP, M. & CHARTER, S. (2001) The CalCOFI
ichthyoplankton time series: Potential contributions to the management of rocky-
shore fishes. California Cooperative Fisheries Investigations, Reports, 42, 112-128.
MYERS, R. (1998) When do environmental-recruitment correlations work? Reviews in Fish
Biology and Fisheries, 8, 285-305.
NEWMAN, S., STECKIS, R., EDMONDS, J. & LLOYD, J. (2000) Stock structure of the
goldband snapper Pristipomoides multidens (Pisces: Lutjanidae) from the waters of
northern and western Australia by stable isotope ratio analysis of sagittal otolith
carbonate. Marine Ecology Progress Series, 198.
NEZLIN, N. & MCWILLIAMS, J. (2003) Satellite data, Empirical Orthogonal Functions,
and the 1997-1998 El Niño off California. Remote Sensing of Environment, 84, 234-
254.
NISHIMOTO, M. & WASHBURN, L. (2002) Patterns of coastal eddy circulation and
abundace of pelagic juvenile fish in the Santa Barbara Channel, California, USA.
Marine Ecology Progress Series, 241, 183-199.
PETERSON, W. & SCHWING, F. (2003) A new climate regime in northeast Pacific
systems. Geophysical Research Letters, 30, doi:10.1029/2003GL017528.
ROY, C., CURY, P. & KIFANI, S. (1992) Pelagic fish recruitment success and reproductive
strategy in upwelling areas: environmental comprimises. South African Journal of
Marine Science, 12, 135-146.
SIEGEL, D., KINLAN, B., GAYLORD, B. & GAINES, S. (2003) Langrangian Descriptions
of Marine Larval Dispersion. Marine Ecology Progress Series, 260, 83-96.
SIRABELLA, P., GUILIANI, A., COLOSIMO, A. & DIPPNER, J. (2001) Breaking down
the climate effects on cod recruitment by principal component analysis and
canonical correlation. Marine Ecology Progress Series, 216, 213-222.
STEELE, M., MALONE, J., FINDLAY, A., CARR, M. & FORRESTER, G. (2002) A
simple method for estimating larval supply in reef fishes and a preliminary test of
population limitation by larval delivery in the kelp bass, Paralabrax clathratus.
Marine Ecology Progress Series, 235, 195-203.
THOMSON, C. (2001) Human Ecosystem Dimension. IN LEET, W., DEWEES, C.,
KLINGBEIL, R. & LARSON, E. (Eds.) California's living marine resources: A
status report., California Fish and Game.
131
Chapter 5: Temporal and spatial genetic variance in the kelp bass, Paralabrax
clathratus
Abstract
Temporal variance is increasingly studied in marine population genetics. Originally
described in the context of unusual spatial structure, temporal variance can be indicative of a
small effective size and a need for non-standard management strategies. We surveyed the
kelp bass, Paralabrax clathratus, for evidence of temporal variance and low effective
population size by collecting 2515 adults and 447 recruits. Sample locations ranged from
the southern side of the Vizcaíno Peninsula, Baja California Sur, Mexico, to Santa Barbara,
California, USA, and represented 31 year classes. Low, yet significant levels of
unpredictable genetic variance were found using a RFLP digest of the mitochondrial control
region and 7 microsatellite loci. Many location-year classes had significant gametic
disequilibrium, despite evidence that the markers are unlinked. Although there was no
indication of sib-ship in the recruit samples, there was evidence of reduced allelic richness.
Variance effective size was found to be 970 (95% C.I., 515: 1566) for the mitochondrial
locus and 4013 (95% C.I., 1684: 7336) for the microsatellite loci. While these results do not
promote management efforts based on the protection of isolation-created diversity, they do
encourage the maintenance of a diversity of breeding populations with large census sizes.
Introduction
The idea that temporal variance is a relevant component of marine population genetics has
been gaining traction with implementation of an increasing number of studies (Hedgecock,
1986, Hedgecock and Sly, 1990, Hedgecock et al., 1992, Hedgecock, 1994b, David et al.,
1997, Kordos and Burton, 1993, Edmands et al., 1996, Pujolar et al., 2006). Chaotic Genetic
Patchiness (CGP), first introduced by Johnson and Black (1982), is an important synoptic
132
explanation for connecting this temporal structure to spatial patterns. The theory’s core
tenets, that marine species can have genetic variance between age classes and that this
variance can express itself as spatial variance, are attributable to the life-cycles of most
marine species. Their high fecundity and mortality rates of dispersive larval stages have
been theorized to cause temporal variance because of genetic drift/small N
e
(Hedgecock,
1994a), transport (Kordos and Burton, 1993) and selection (Johnson and Black, 1982).
Recognizing the existence of temporal variance has a number of implications. First, change
in allele frequencies from generation t to t+1 is reflective of effective population size (N
e
)
according to the relationship:
!
"[(x
t +1
#x
t
)
2
] =
x
t
(1#x
t
)
2N
e
(Pollak, 1983). In other words,
effective population size is the size of an “ideal population” with predictable loss of genetic
variation over time (Crow and Kimura, 1970). Deviations in random mating, equal sex
ratios, binomial- or Poisson-distributed variance in family sizes, non-overlapping
generations and constant population size can reduce N
e
and cause large populations to “act”
genetically much smaller (Hedrick, 2005). While a species with a large effective size might
build up population differences in a slow, step-wise manner (Hartl and Clark, 1989), an
organism with a small N
e
will experience more drift events (Schwartz et al., 1998) and lose
genetic diversity (Hedrick, 2000). This process will also alter the balance between drift and
selection, reducing a population’s response to selective forces (see Hedgecock, 1994a).
Second, temporal variance has management consequences for fishery species. Were the
production of recruits variable, management would need to annually adjust fishing pressure.
If variance is unpredictable, as implied by sweepstakes recruitment or transport, then bet
hedging reserves would be a useful strategy (Larson and Julian, 1999).
133
One limitation in the study of temporal variance has been the ability to define adult cohorts.
Comparisons have been made between repeated samples of adult populations (e.g., Lacson
and Morizot, 1991, Hedgecock et al., 1994), between adult populations and settling recruits
(e.g., Edmands et al., 1996, Johnson and Wernham, 1999, Hedgecock et al., 2006) or
between waves and/or years of the recruits themselves (e.g., Li and Hedgecock, 1998,
Pujolar et al., 2006). Although size can be occasionally used (Moberg and Burton, 2000),
variance in growth rates has made dividing adult populations more problematic.
Alternatively, though, adult fish may be segregated into cohorts because of three pairs of
bones in the brain cavity, called otoliths, which collect sensory inputs of sound, gravity,
movement and acceleration (Popper et al., 2005). While not easily discernable for all
species, daily and seasonal banding occurs with cyclical accretion of calcium and protein
matrix to the metabolically inert otoliths (Helfman et al., 1997). Birth dates or years of
individual fish can be calculated through the scoring of this banding and has been employed
to analyze temporal variance in white sea bream (Diplodus sargus) (Lenfant and Planes,
2002), brown trout (Salmo trutta) (Jorde and Ryman, 1996), haddock (Melanogrammus
aeglefinus) (Purcell et al., 1996) and striped bass (Morone saxatilis) (Chapman, 1989).
We report here the analysis of temporal and spatial genetic variance in kelp bass, Paralabrax
clathratus, a serranid abundant in kelp and rock structure along the southern California and
Baja coast (Eschmeyer et al., 1983, Love, 1996). It is an important component of
California’s $300 million dollar sport fishing industry (Thomson, 2001) as it is one of the
most frequently caught species in the southern California region (Love, 1996).
134
Beyond contribution to management, we studied this species for a number of reasons. First,
kelp bass can live for 30 years (Love, 1996), allowing collection of a variety of year classes.
Second, because the species is regularly caught on sport fishing boats, year classes could be
heavily sampled. Third, otolith scoring has been previously performed (Love et al., 1996).
The analysis was dependent on confidence in the ageing method, which cannot be achieved
in all species because of obstructive growth patterns or non-seasonal checking. Unlike
bocaccio rockfish (Sebastes paucispinis) for example (Andrews et al., 2005), Paralabrax
clathratus has clear banding, confirmed to be annual in nature (Love et al., 1996).
Oxytetracycline has also been used to confirm the first winter band (Vogel et al., in prep-a).
Fourth, analyzing kelp bass, which have a long lifespan (Love, 1996), large abundance
(Allen and Hovey, 2001) and high fecundity (Oda et al., 1993), would be a salient indicator
of the importance of recognizing the existence of temporal variance. Although a species
such as this would not immediately be expected to show evidence of temporal variance,
studies have shown that fecund broadcast spawners such as Paralabrax clathratus can have
extremely low
!
N
e
N
ratios (Hauser et al., 2002, Turner et al., 2002, Hoarau et al., 2005).
Finally, understanding temporal variance would help explain recruitment and be a foil for
other studies. Kelp bass recruitment has been correlated to kelp frond density (Macrocystis
pyrifera) (Carr, 1994) and tidal bores (Findlay and Allen, 2002); otolith samples used here
connected recruitment to Triad Zones of enrichment, concentration and retention (Vogel et
al., in prep-a). Preliminary studies have also been performed on micronutrient signatures (J.
Casselle, pers. comm.) in a manner similar to the analysis of Thallassoma bifasciatum (blue-
headed wrasse) (Swearer et al., 1999), which found evidence of self recruitment patterns.
135
We tested Paralabrax clathratus for evidence of spatial and temporal genetic variance.
Recruits were also analyzed for allelic richness, which has been proposed to be low for
species with high variance in reproductive output and small effective population size
(Hedgecock, 1994a, Hedgecock et al., 2006). These tests were complimented with analysis
of relatedness within recruit samples and estimation of variance N
e
for the entire species.
Materials and Methods
Sampling Strategy
We collected 2515 adults across 6 locations from 2001 to 2003 (Figure 22, Table 19). Fish
caught at the Abreojos (AB) location were collected with plastic lures; most samples from
the other locations were taken from leftover carcasses of sport fishing cruises. These fish
were primarily caught with live anchovy and sardine.
Recruit samples were collected by T. Anderson on Santa Catalina Island in 2004 (148
individuals) and 2005 (100 individuals). These were augmented by a selection of recruits
collected in 1998 (100 individuals) (Findlay and Allen, 2002); another non overlapping part
of that sample has been previously reported in (Luzier and Wilson, 2004). A final sample of
recruits (99 individuals) collected in Carpenteria, CA (just south of Point Conception, Figure
22) in 2001 were provided by J. Casselle and K. Selkoe (total recruit sample size = 447).
The 2004 and 2005 S. Catalina samples were collected by sweeping kelp stands with fine-
mesh netting, while the others were taken from artificial recruitment collectors described in
(Ammann, 2004). All four recruit samples represented multiple recruitment events in that
there was a large range in total length (60-340mm).
136
Table 19. Sampling locations names (marked with letters as in Figure 22), sample sizes (n),
number of year classes represented in each sample, the range of those year classes, sample
site coordinates, and sample dates. Sample SCR98 was collected by (Findlay and Allen,
2002); sample CPR01 was provided by J. Caselle and K. Selkoe.
137
Figure 22. Sampling locations of Paralabrax clathratus: (AB) Abreojos, Baja California
Sur, MX; (CO) Coronado Islands, Baja California Norte, MX; (SD) San Diego, California,
USA; (CE) San Clemente Island, California, USA; (SC, SCR98/SCR04/SCR05 for recruits)
Santa Catalina Island, California, USA; (CPR01 for recruits) Carpenteria, California, USA;
(SA) Santa Barbara, California, USA.
138
Scoring of Otoliths
Otolith scoring is reported in (Vogel et al., in prep-a). Briefly, otoliths were sectioned with a
low-speed Buehler Isomet Saw® and using a 25X trans-illuminating microscope, only the
clearest samples were saved. Confirmation of the first winter band (winter banding
confirmed by Love et al., 1996) was accomplished with oxytetracycline marking of three
young-of-the-year. Otolith scoring was performed independently by two readers. If
disagreement could not be solved by another reading, the second otolith was sectioned. Ten
samples were removed from temporal analysis because of unscoreable patterns.
Genetic Markers
One mitochondrial and seven microsatellite markers were utilized in the analysis. The
mitochondrial locus was scored with Restriction Fragment Length Polymorphisms (RFLPs)
of the control region, the method of which is described in (Vogel et al., in prep-b). We used
four enzymes (AciI, AflIII, DdeI and !TaqI) to cut PCR product amplified with Primers “K”
of Lee et al. (1995) and “12RH” of Sivasundar et al. (2001) (the use of AflIII requires “K”
and an alternate primer “233Cutter” (Vogel et al., in prep-b). Banding was scored using
“Spreadex” acrylimide gels (www.elchrom.com) stained with Ethidium Bromide and
recorded with a Kodak Gel Imaging System Version 2.0.2.
The seven microsatellites (AV6, AV15, AV17, AV88, AV115, Gag010 and Mbo66) are
reported in Selkoe et al. (2005) and Vogel et al. (in prep-b). Fragments were separated and
scored on a Beckman CEQ 8000 sequencer and version 8.0 of its software. Because of
amplification problems, one of the recruit samples, SCR98, was not scored.
139
Analyses Performed
Because age structure was stable across the 2001 and 2002/3 sample seasons (Vogel et al., in
prep-a), all tests were performed on the combined data. We found some differences in
pairwise F
ST
estimates, indicating the locations were not perfectly genetically identical
between sample seasons (data not shown). All estimates though, were extremely low;
pairwise !
ST
values averaged 0.0019 (±0.0007) and pairwise " values 0.0007 (±0.0002).
Estimates as low as these are prone to error but are improved by repetitive sampling
(Waples, 1998, Vogel and Edmands, in prep-a). By combining the two seasons, we in effect
averaged the values for a more accurate estimation.
Summary statistics were collected with Arlequin v.3.01 (Excoffier et al., 2005) and Genetix
(Belkhir et al., 2004) for data grouped by adult location, recruit sample and location-year
class. Allele number, frequency and diversity were collected for both types of loci;
microsatellites were tested with Genetix for significant F
IS
and Linkage Disequilibrium (LD)
with 500 permutations. Estimates of F
IS
over all loci are reported. With seven microsatellite
loci, there were 21 pairwise comparisons of LD within each group. The analysis of a group
was considered significant if more than 5% of the pairwise comparisons were significant.
Global tests of genetic heterogeneity for combined microsatellite and mitochondrial data
were performed on samples grouped as locations and location-year class with an estimation
of " (Weir and Cockerham, 1984) (significance determined with 1000 permutations).
Pairwise associations were tested with 1000 permutations and visualized with a Principal
Components Analysis (PCA) of log-ratio transformed (Aitchison, 2003) allele frequencies.
140
Recruit samples were tested for low allelic richness with regards to a sweepstakes prediction
(Hedgecock, 1994a). Because sample size influences the number of alleles recorded, we
compared the number of alleles in a recruit sample versus the mean number found in
equivalently-sized, random adult samples. For each sample size, a distribution of 10,000
allelic counts was created based on random adult samples with the same size as the recruit
sample. If 95% of the 10,000 random samples had a greater number of alleles than the
recruit sample, allelic richness of the recruit sample was significantly low.
Analysis of sib-ship was performed with Kinship (Goodnight and Queller, 1999). Recruit
samples was tested for full (r=0.50) and half (r=0.25) sib-ship with significance determined
by comparison to 10,000 simulated pairs. The frequency of significant pairs beyond # (0.05,
0.01, 0.001) was used to determine if there were a high number of close relationships.
Lastly, N
e
was estimated for the species. To reduce sampling error, analysis included only
the large year classes from 1992 to 1997. Using a life table and estimation of fecundities
based on the results of Oda et al. (1993), generational length (G) and a correction factor for
overlapping generations (C) were calculated in “factor” (supplied by P. E. Jorde) (Jorde and
Ryman, 1995). The estimate of allele frequency shift,
!
" F (Pollak, 1983), between
consecutive cohorts was adjusted according to
!
ˆ
N
e
=
C
2G " F
(microsatellites) and
!
ˆ
N
e
=
C
G " F
(mitochondrial marker) (Jorde and Ryman, 1995). As suggested by (Jorde and Ryman,
1995), two confidence intervals were created based on a $
2
distribution for n alleles (Waples,
1989)
!
(n"1) # F
$/2
2
%
[df = n"1]
,
(n"1) # F
1"$/2
2
%
[df = n"1]
&
'
(
(
)
*
+
+
and standard error of a normal distribution.
141
Results
Summary statistics, F
IS
and linkage disequilibrium
Sample sizes, number of alleles and allelic diversity/heterozygosity for each location and
location-year class are reported in Appendices 1 and 2. As expected for a control region
marker and microsatellites, allele diversities and heterozygosity were generally high,
although there was considerable variation among loci (Table 20).
Two location samples (AB and SD) exhibited significant positive F
IS
(Table 21). Locus by
locus analysis found contrasting patterns; all loci for AB had positive F
IS
, whereas F
IS
in the
SD sample was driven by the Gag010 microsatellite (data not shown). In tests of adult
locations and recruit samples with Micro-Checker (Oosterhout et al., 2004), three of 63 tests
(4.8%) found evidence of null alleles, below the error rate of 5%. Each significant result
was from different markers at different locations. No large allele dropout was found.
Patterns of LD were more broadly distributed than F
IS
(Table 22) but again, no particular loci
dominated results. Gametic disequilbrium was found in 16/100 year classes from all
locations except SD, but was concentrated in the AB (7/16) and CE (4/16) samples. One
recruit sample (SCR05) had high LD. Location samples (3/6) confirmed the high LD in the
AB location-year classes but also indicated strong LD in the SC and CO collections.
Spatial structure
Global spatial structure was extremely low, but significant (" = 0.0005, 95% CI: 0.00007 -
0.00113, p = 0.002). Differences between locations were unpredictable (Figure 23) and
results of microsatellite data did not match mitochondrial data (Table 23). Structure seen in
142
Table 20. Global summary statistics for RFLP analysis of the mitochondrial control region
maker and seven microsatellites.
Locus h/H # of Alleles Primer Reference
mt Control Region 0.747 42 K: Lee et al., 1995;
12RH: Sivasundar et al., 2001
AV6 0.952 62 Selkoe et al., 2005
AV15 0.918 38 Selkoe et al., 2005
AV17 0.603 23 Selkoe et al., 2005
AV88 0.927 27 Selkoe et al., 2005
AV115 0.877 22 Selkoe et al., 2005;
Vogel et al., in prep-b
Gag010 0.773 21 Selkoe et al., 2005
Mbo66 0.406 3 Selkoe et al., 2005
143
Table 21. F
IS
scores for locations and location-year classes. Columns refer to locations in
Figure 22. Data grouped by location (including recruit samples) are in the row “Loc.”
Areas of no data indicate location-year classes that were not collected. *Significant values
of F
IS
are underlined and in italics.
AB CO SD CE SC SB SCR04 SCR05 CPA01
Loc. 0.012* 0.005 0.013* -0.005 0.008 -0.020 0.002 0.032* -0.003
1969 0.000
1973 0.000
1974 0.000
1975 0.000 0.000
1976 0.080
1977 -0.091
1979 -0.078 0.000 -0.250
1980 -0.029 0.186*
1981 0.013 0.099*
1982 -0.041 -0.182 -0.028
1983 -0.108 -0.097
1984 -0.038 0.000 -0.024
1985 0.082* 0.013 0.000 -0.020
1986 0.064 0.000 0.091 -0.069 0.000
1987 -0.014 -0.083 0.000 0.026
1988 -0.044 0.245* 0.000 0.027 0.000
1989 0.017 0.046 -0.044 -0.030 0.087
1990 0.021 0.070 0.004 0.000 0.231*
1991 -0.029 0.029 0.027 -0.011 0.115* 0.084
1992 0.026* -0.030 0.035 0.016 0.015 -0.075
1993 -0.023 0.022 0.014 0.038 0.030 -0.016
1994 0.013 -0.017 0.029* -0.026 -0.065 -0.021
1995 0.011 -0.001 0.031* -0.004 -0.025 -0.061
1996 0.022* 0.050* -0.008 0.052 0.029 0.013
1997 0.012 -0.012 -0.022 -0.116 -0.009 -0.022
1998 0.034 -0.053 0.083 -0.100*
1999 0.013 0.000 0.152*
2000 0.011
144
Table 22. Percentage of significant LD scores for locations and location-year classes.
Columns refer to locations in Figure 22. Data grouped by location (including recruit
samples) are in the row “Loc.” Areas of no data indicate location-year classes that were not
collected. *Percentages of LD that are underlined and in italics are greater than 0.05 and
significant.
AB CO SD CE SC SB SCR04 SCR05 CPA01
Loc. 0.286* 0.095* 0.000 0.048 0.095* 0.048 0.048 0.143* 0.048
1969 0.000
1973 0.000
1974 0.000
1975 0.000 0.000
1976 0.000
1977 0.000
1979 0.000 0.000 0.000
1980 0.095* 0.000
1981 0.048 0.000
1982 0.048 0.000 0.000
1983 0.000 0.000
1984 0.191* 0.000 0.048
1985 0.000 0.048 0.000 0.000
1986 0.000 0.000 0.000 0.048 0.000
1987 0.048 0.000 0.000 0.095*
1988 0.048 0.048 0.000 0.095* 0.000
1989 0.048 0.000 0.000 0.000 0.000
1990 0.238* 0.000 0.048 0.000 0.000
1991 0.095* 0.048 0.048 0.000 0.000 0.000
1992 0.143* 0.000 0.000 0.095* 0.048 0.000
1993 0.000 0.000 0.000 0.000 0.191* 0.048
1994 0.000 0.000 0.048 0.048 0.048 0.095*
1995 0.143* 0.000 0.048 0.095* 0.000 0.000
1996 0.048 0.095* 0.048 0.000 0.048 0.000
1997 0.095* 0.000 0.000 0.000 0.048 0.000
1998 0.048 0.095* 0.000 0.095*
1999 0.000 0.000 0.000
2000 0.048
145
Table 23. Pairwise F
ST
estimates for location data. ID’s refer to locations in Figure 22.
Values above the diagonal describe microsatellite data and are " (Weir and Cockerham,
1984), below describe mitochondrial data and are !
ST
(Excoffier et al., 1992). *Values
underlined and in italics are significant.
AB CO SD CE SC SB SCR04 SCR05 CPR01
AB 0.0003 -0.0002 0.0003 0.0001 0.0014* -0.0001 0.0034* 0.0004
CO 0.0102 * 0.0023 -0.0001 0.0006 0.0009* 0.0001 0.0019* -0.0001
SD 0.0034 * 0.0001 0.0001 -0.0002 0.0009 -0.0002 0.0026* 0.0002
CE -0.0003 0.0059* 0.0042* 0.0008 0.0013* -0.0003 0.0031* 0.0007
SC 0.0001 0.0035 0.0009 -0.0010 0.0006 0.0006 0.0022* 0.0015
SB -0.0016 0.0100* 0.0004 0.0017 -0.0011 0.0014 0.0008 0.0011
SCR04 -0.0035 0.0002 -0.0045 0.0041 -0.0053 -0.0051 0.0042* 0.0005
SCR05 -0.0033* 0.0034 -0.0001 -0.0039 -0.0044 -0.0030 -0.0075 0.0074
CPR01 0.0022 0.0034 0.0101 0.0128 0.0064 0.0197* 0.0041 0.0013
146
Figure 23. Plot of Component 1 versus Component 2 from Principal Components Analysis
of location and temporal data (plus recruit samples). Location ID’s correspond to Figure 22.
Eigenvalues for component 1 were 2.953 (25.0% of variance) for location data and 1.019
(14.4%) for location-year class data. Eigenvalues for component 2 were 2.145 (18.1% of
variance) for location data and 0.779 (11.0%) for location-year class data.
147
the mitochondrial data was in large part defined by significant differences with the AB, CO,
SD, and CE locations (below diagonal, Table 23). The CPR01 recruit sample was
differentiated from the nearby adult sample of SB. Structure in the microsatellite data was
pronounced for the SB and the SCR05 recruit sample (above diagonal, Table 23).
Temporal structure
Global structure by location-year class was non-significant (" = 0.0002, p = 0.319). The
PCA plot of location-year classes supported this result in that many of the year classes were
tightly clustered (Figure 23). This plot also indicated that the spatial structure was not
consistent across year classes, in that location-year classes did not group by location.
We further analyzed year class influence by plotting the effect of removing single years on
spatial differentiation (Figure 24). Absolute deviations from the total sample estimates were
positively correlated to the size of the year class removed (Figure 24; AB: R
2
= 0.2407, p =
0.0109, SD: R
2
= 0.4205, p = 0.0089; result is consistent across other comparisons [data not
shown]). The relative effect though, was randomly distributed across year classes (Figure
24; AB: R
2
= 0.0.0603, p = 0.7700, AB: R
2
= 0.0183, p = 0.6310). Were spatial structure
stable, the second comparison would show evidence of positive trend.
A final note concerns recruit samples. Unlike the comparisons to locations, recruit samples
grouped better in that none were isolated in the principal component space (Figure 23). This
result indicated that when compared to adult cohorts, recruit samples were not particularly
unique (e.g., (Moberg and Burton, 2000). Only viewed against adult location samples,
which are an amalgamation of year classes, did the recruit samples begin to appear unusual.
148
Figure 24. Plots of the influence on the !
ST
estimate between Abreojos (AB) and San Diego
(SD) with removal of individual location-year classes. Estimates are plotted along the x-axis
by the log
10
value of the number of individuals removed. The dashed line corresponds to the
!
ST
estimate for the two samples without removal of any year class.
149
Allelic richness
One of the three tests for reduced allelic richness was significant to the Bonferroni
Correction level of 0.0125 (Table 24). The SCR98 sample was not tested because of no
microsatellite scores. It is interesting to note that all of the samples had low, if not
significant, diversity.
Relatedness
Although there were significant sib-ship associations between individual recruits,
frequencies never rose above the error limits of 0.05, 0.01 and 0.001 (data not shown).
Considering the evidence of reduced allelic richness and low N
e
(below), it would have been
reasonable to find sibling-infused recruit classes. Our recruit samples were not from single
waves of arrival though, reducing collection of sibling cohorts.
Effective population size
Mean generation time (G) was 6.6 years and the correction factor (C) was 35.25. The
temporal effective population size for the species was 970 for the mitochondrial locus and
4013 for the microsatellite loci (Table 25), a pattern that is expected from mitochondrial loci
being haploid and maternally inherited. Both types of confidence intervals ($
2
and normal)
differentiated the estimates from %; the more conservative $
2
interval ranged from 516 to
1566 for the mitochondrial locus and 1684 to 7336 for the microsatellite loci.
150
Table 24. Results from testing for deficiency of alleles in the recruit samples. ID’s refer to
Table 19 and Figure 22. Simulated richness (a
s
) is the mean number of alleles (including
both mitochondrial and microsatellite alleles) found with 10,000 n-sized random samples of
adult data. Simulation was adjusted to account for the difference in n between mitochondrial
and microsatellite data. **significant after sequential Bonferonni correction.
Mitochondrial Locus Microsatellite Loci
n # alleles n # alleles a
s
p value
CPR01 86 15 94 114 136 0.0549
SCR04 99 18 143 122 147 0.0464
SCR05 99 13 94 114 137 0.0132**
151
Table 25.
!
" F , its 95% confidence interval based on both a $
2
distribution and a normal
distribution and the corresponding values of
!
ˆ
N
e
with adjustments by generation length (G)
and a correction factor for overlapping generations (C). Samples from 1992 to 1997 were
analyzed.
Mitochondrial Locus Microsatellite Loci
!
" F 0.0055 0.0027
C.I. ($
2
distribution) (0.104, 0.0034) (0.0063, 0.0015)
C.I. (normal distribution) (0.0073, 0.0037) (0.0032, 0.0021)
!
ˆ
N
e
970 4013
C.I. ($
2
distribution) (516, 1566) (1684, 7336)
C.I. (normal distribution) (735, 1429) (3296, 5127)
152
Discussion
F
IS
and Linkage Disequilbrium
Although there was evidence of significant F
IS
, there was no indication of null alleles or
large allele dropout. There was also LD in the samples, but during analysis of Mendelian
inheritance, Vogel et al. (in prep-b) found no significant linkage. An interesting possibility
centers on F
IS
and LD being confined to certain cohorts. Hauser et al. 2002) connect F
IS
to
inbreeding in heavily-fished populations with small effective population sizes. Gametic
disequilibrium, if not caused by linkage, can be reflective of recent genetic drift (LeBerg,
2005). While we did not assess N
e
with LD data because of sample sizes below 90-100
(Bartley et al., 1992), we consider the patterns evident of variability between cohorts.
Temporal variance in the marine environment
We have shown evidence that Paralabrax clathratus exhibits temporal variability. There
was LD in a significant number of locations and location-year classes, with no evidence of
linkage (Vogel et al., in prep-b). Sample locations had significant (and low), unexpected
affinities and location-year classes did not appear to cluster by location. Larger location-
year classes, while influencing the differentiation estimate because of increased size, did not
typically increase population differentiation, as might be expected with a stable location
signal. Recruit samples grouped with location-year classes but typically clustered away
from adult location samples. One recruit sample had low allelic richness and the other two
samples tested, although not significant, also had tantalizingly low numbers of alleles. An
estimation of temporal N
e
found bounded values of 970 for the mitochondrial locus and 4013
for the microsatellite loci. This result corresponded to expected differences between locus
types, and is a fraction of the census size, likely in the tens of millions.
153
These final two results were based on suggested tests for evidence of high reproductive
variance (Hedgecock, 1994a). Another possible cause, selection, is impossible to confirm or
rule out, especially with only eight markers. Perhaps as more functional gene assays become
available, surveys of genes important for “larvae-important” traits can be performed.
Whether transport is a the single cause is also a difficult question. Paralabrax clathratus is
highly fecund (Oda et al., 1993) with minimal differentiation and undoubtedly long distance
larval dispersal. There is reason to believe however, that locations are not regularly bathed
with recruits from across the species’ range. Although recruitment showed similarities
between four of the locations, there were localized patterns for all sample sites (Vogel et al.,
in prep-a). Larvae collected in California Cooperative Oceanic Fisheries Investigations
(CalCOFI) plankton tows are also almost exclusively found in near-shore waters (Moser et
al., 2001), in close proximity to adult populations. These patterns show that when transport
takes place, it does not equally benefit all locations. For it to be a driver of the regular
temporal variance, it would need to consistently occur but be variable in its arrival sites.
Effective size and its ratio with N
Kelp bass have a small effective population size and are not expected to behave genetically
as might be implied by their census size (Hedrick, 2005). Estimating that census numbers
are between 5 and 50 million fish (with an even division between male and female), the
!
N
e
N
ratio is between 2&10
-4
and 2&10
-5
for the mitochondrial genome and between 8&10
-4
and
8&10
-5
for the nuclear genome. These values are equivalent to a number high fecundity
species with smaller than expected effective sizes (Hauser et al., 2002, Hoarau et al., 2005).
154
We can only speculate about the causes of these low ratios. The values are consistent with
the shallow phylogenetic relationships in the control region for Parlabrax clathratus (Vogel
and Edmands, in prep-b), a pattern proposed to be indicative of variability in reproductive
success (Hedgecock, 1994a). Beyond low reproductive success, proposals have included
immigration of new alleles (Nunney, 1996), variable production of demes (Turner et al.,
2002) and skewed male success (Bekkevold et al., 2002). Skewed male success appears to
be a candidate, because of spawning aggregations. Erisman and Allen (2006) recorded 20
different behaviors involved with courtship and spawning; in general, gravid females are
berated by clusters of males. This style of courtship could limit the number of successful
spawners by selecting for the most active or dominant individuals.
Relationship to other analyses of recruitment
Our interpretation of the relationship of CalCOFI survey results (Moser et al., 2001) to our
data has already been presented, but Carr (1994) and Findlay and Allen (2002) also
described environmental events that encourage recruitment. What remains unclear is the fate
of larvae that do not gain access to these events. Our data support the established idea that
mortality in the plankton is intense. Although only suggestive, this would indicate that kelp
frond density (Carr, 1994) or tidal bores (Findlay and Allen, 2002) may not just be
recruitment cues or methods, but rather highly variable survival opportunities.
Recruitment has also been correlated to “Triad” events (Vogel et al., in prep-a). These
oceanic patterns, previously only associated in the areas inhabited by kelp bass with
California northern anchovy (Engraulis mordax) (Cury et al., 1995), occur in the Southern
California Bight and the Gulf of Ulloa (see Figure 22) (Bakun, 1996). “Enrichment,” or
155
upwelling, is important for larval survival, but can advect larvae away from proper habitat or
cause too much water column turbulence. “Retention” and “concentration” of larvae and
nutrients are therefore components of recruitment success. As anticipated by the theory,
recruitment of Paralabrax clathratus was highest for moderate levels of upwelling/retention.
An interesting connection to that study is the evidence that year class impact on the spatial
signal increased with sample size, although direction did not. This suggests that the cause of
temporal variance is independent of recruitment success and variability is a consistent part of
the Paralabrax clathratus life cycle. If variable spawning success is the primary driver for
example, then even in successful years, such a small percentage of larvae will survive that
another variance event will occur. On the other hand, if variance is driven by breeding
behavior, behavior won’t change as drastically as the variability seen in recruitment.
Anecdotally for example, Erisman and Allen (2006) report spawning behavior in the summer
of 2002, despite that year being a recruitment failure for the species (pers. obs.; also seen in
survey data from the Kelp Forest Monitoring Program [Vogel et al., in prep-a]).
Management implications
The kelp bass fishery is large; 1.7 million individuals per year were removed on average
from California waters in the 1990’s (Allen and Hovey, 2001). The California Department
of Fish and Game (CDFG) does not individually manage the species, but organizes it with
two sister species, Paralabrax nebulifer (barred sand bass) and P. maculatofasciatus (spotted
sand bass). In any one day, no more than 10 Paralabrax spp. may be kept, all must be 12
inches or larger and their fillets must be at least 6.5 inches (Kellog et al., 2006).
156
Most immediately, we have not shown evidence of isolated populations. While the age
structure data imply localized character, there is no argument to be made about protecting
the genetic diversity found in unique populations. The subtle structure reported here was a
reflection of small allele frequency differences and not unique complements of alleles.
The small
!
N
e
N
ratio presents a number of considerations. While effective size was not so
small that heterozygosity was low (marine fish have microsatellite heterozygosities of
0.79±0.26 [DeWoody and Avise, 2000]), it suggests that effects could occur at large census
sizes. Franklin (1980) and Soulé (1980) advocated effective sizes greater than 500 for
maintaining genetic diversity. Based on evidence of higher than expected rates of
detrimental mutation, Lande (1995) proposed that an N
e
of 5000 was a more appropriate
conservation goal. If this larger calculation is accurate, then Paralabrax clathratus is
already potentially impacted and managers should work to keep census sizes as high as
possible (also see Hauser et al., 2002). Social structure and non-random mating behavior
because of fishing pressure have also been theorized to exacerbate the effects of a small
!
N
e
N
ratio and promote inbreeding and genetic drift (Hoarau et al., 2005). While kelp bass do not
have have strong philopatric tendencies (although they use home ranges [Lowe and Topping,
2003]), they do not mate randomly (Erisman and Allen, 2006). This will only increase the
possibility of detrimental genetic effects in higher census sizes than might be anticipated.
Larson and Julian (1999) proposed that evidence of temporal variance should promote a
strategy of bet-hedging reserves because of implied randomness in the breeding population.
Turner et al. (2002) made a similar point for Sciaenops ocellatus (red drum), which they
157
theorized to be affected by variable productivity from demes that use a diversity of estuaries.
Considering the low effective size and the localized character of the age distribution, this
viewpoint could be relevant to kelp bass. Although genetic diversity may not need to be
widely protected, a diversity of viable breeding populations might. Also relevant is the
evidence that maternal age affects larval quality (Berkeley et al., 2004). While more
deterministic in nature than the stochastic reproductive success implied by a small
!
N
e
N
ratio,
it would play an important role if variance is driven by reproductive success and spawning
behavior. We therefore encourage the efforts of many sport fishermen to return large
individuals as well as support the CDFG as they implement a series of coastal reserves.
Acknowledgements
Otolith work was performed in the laboratory of O. Sosa-Nishisaki and J. Rosales-Cassein at
the Centro de Investigación Científica y de Educación Superior de Ensenada (CICESE) in
Ensenada, Baja California, MX Funding for this research was provided by a fellowship from
the National Security Education Program, Rose Hills Summer Fellowships from the Wrigley
Institute for Environmental Studies and other funding from the Wrigley Institute for
Environmental Studies.
158
Chapter 5 References
AITCHISON, J. (2003) The statistical analysis of compositional data, Caldwell, NJ,
Blackburn Press.
ALLEN, L. & HOVEY, T. (2001) Kelp bass. IN LEET, W., DEWEES, C., KLINGBEIL, R.
& LARSON, E. (Eds.) California's living marine resource: a status report.
California Department of Fish and Game.
AMMANN, A. (2004) SMURFs: standard monitoring units for the recruitment of temperate
reef fishes. Journal of Experimental Biology and Ecology, 299, 135-154.
ANDREWS, A., BURTON, E., KERR, L., CAILLIET, G., COALE, K., LUNDSTROM, C.
& BROWN, T. (2005) Bomb radiocarbon and lead-radium desequilibria in otoliths
of bocaccio rockfish (Sebastes paucispinis): a determination of age and longevity for
a difficult-to-age fish. Marine and Freshwater Research, 56, 517-528.
BAKUN, A. (1996) Patterns in the Ocean, Ocean Processes and Marine Population
Dynamics, La Paz, BCS, Mexico, California Sea Grant/CIB.
BARTLEY, D., BAGLEY, M., GALL, G. & BENTLEY, B. (1992) Use of linkage
disequilibrium data to estimate effective size of hatchery and natural fish
populations. Conservation Biology, 6, 365-375.
BEKKEVOLD, D., HANSEN, M. & LOESCHCKE, V. (2002) Male reproductive
competition in spawning aggregations of cod (Gadus morhua, L.). Molecular
Ecology, 11, 91-102.
BELKHIR, K., BORSA, P., CHIKHI, L., RAUFASTE, N. & BONHOMME, F. (2004)
Genetix 4.05, logiciel sous Windows TM pour la génétique des populations. IN
LABORATOIRE GÉNOME, P., INTERACTIONS, CNRS UMR 5171 (Ed.)
Montpellier (France), Université de Montpellier II.
BERKELEY, S., CHAPMAN, C. & SOGARD, S. (2004) Maternal age as a determinant of
larval growth and survival in a marine fish, Sebastes melanops. Ecology, 85, 1258-
1264.
CARR, M. (1994) Effects of macroalgal dynamics on recruitment of a temperate reef fish.
Ecology, 75, 1320-1333.
CHAPMAN, R. W. (1989) Spatial and temporal variation of mitochondrial DNA haplotype
frequences in the striped bass (Morone saxatilis) 1982 year class. Copeia, 2, 344-
348.
CROW, J. & KIMURA, M. (1970) An introduction to population genetics theory,
Minneapolis, Burgess.
CURY, P., ROY, C., MENDESSOHN, R., BAKUN, A., HUSBY, D. & PARRISH, R.
(1995) Moderate is better: exploring nonlinear climatic effects on the Californian
northern anchovy (Engraulis mordax). Canadian Journal of Fisheries and Aquatic
Science, 121, 417-424.
DAVID, P., PERDIEU, M., PERNOT, A. & JARNE, P. (1997) Fine-grained spatial and
temporal population genetic structure in the marine bivalve Spisula ovalis.
Evolution, 51, 1318-1322.
DEWOODY, J. & AVISE, J. (2000) Microsatellite variation in marine, freshwater and
anadromous fishes compared with other animals. Journal of Fish Biology, 56, 461-
473.
EDMANDS, S., MOBERG, P. & BURTON, R. (1996) Allozyme and mitochondrial DNA
evidence of population subdivision in the purple sea urchin Strongylocentrotus
purpuratus. Marine Biology, 126, 443-450.
159
ERISMAN, B. & ALLEN, L. (2006) Reproductive behaviour of a temperate serranid fish,
Paralabrax clathratus (Girard), from Santa Catalina Island, California, USA.
Journal of Fish Biology, 68, 157-184.
ESCHMEYER, W. N., HERALD, E. S. & HAMMANN, H. (1983) A Field Guide to Pacific
Coast Fishes, New York, New York, Houghton Mifflin Company.
EXCOFFIER, L., LAVAL, L. & SCHNEIDER, S. (2005) Arlequin ver. 3.0: An integrated
software package for population genetics data analysis. Evolutionary Bioinformatics
Online, 1, 47-50.
FINDLAY, A. & ALLEN, L. G. (2002) Temporal patterns of settlement in the temperate
reef fish Paralabrax clathratus. Marine Ecology Progress Series, 238, 237-248.
FRANKLIN, I. (1980) Evolutionary changes in small populations. IN SOULÉ, M. &
WILCOX, B. (Eds.) Conservaion biology: An evolutionary-ecological perspective.
Sunderland, MA, Sinauer Associates.
GOODNIGHT, K. & QUELLER, D. (1999) Computer software for performing likelihood
tests of pedigree relationship using genetic markers. Molecular Ecology, 8, 1231-
1234.
HARTL, D. & CLARK, A. (1989) Principals of Population Genetics, Sunderland, Sinauer
Associates, Inc.
HAUSER, L., ADCOCK, G., SMITH, P., BERNAL RAMÍREZ, J. & CARVALHO, G.
(2002) Loss of microsatellite diversity and low effective population size in an
overexploited population of New Zealand snapper (Pagrus auratus). Proceedings of
the National Academy of Sciences of the United States of America, 99.
HEDGECOCK, D. (1986) Is gene flow from pelagic larval dispersal important in the
adaptation and evolution of marine invertebrates? Bulletin of Marine Science, 39,
550-564.
HEDGECOCK, D. (1994a) Does variance in reproductive success limit effective population
sizes of marine organisms? IN BEAUMONT, A. (Ed.) Genetics and Evolution of
Aquatic Organisms. London, Chapman and Hall.
HEDGECOCK, D. (1994b) Temporal and spatial genetic structure of marine animal
populations in the California current. California Cooperative Fisheries
Investigations, Reports, 35, 73-81.
HEDGECOCK, D., CHOW, V. & WAPLES, R. (1992) Effective population numbers of
shellfish broodstocks estimated from temporal variance in allelic frequencies.
Aquaculture, 108, 215-232.
HEDGECOCK, D., HUTCHINSON, E., LI, G., SLY, F. & NELSON, K. (1994) The central
stock of northern anchovy Engraulis mordax is not a randomly mating population.
California Cooperative Fisheries Investigations, Reports, 35, 121-136.
HEDGECOCK, D., LAUNEY, S., PUDOVKIN, A., NACIRI, Y., LAPÈGUE, S. &
BONHOMME, F. (2006) Small effective number of parents (N
b
) inferred for a
naturally spawned cohort of juvenile European flat oysters Ostrea edulis. Marine
Biology, doi: 10.1007/s00227-006-0441-y.
HEDGECOCK, D. & SLY, F. (1990) Genetic drift and effective population sizes of
hatchery-propagated stocks of the Pacific oyster, Crassostrea gigas. Aquaculture,
88, 21-38.
HEDRICK, P. (2000) Genetics of populations, London, Jones and Bartlett Publishers
International.
HEDRICK, P. (2005) Large variance in reproductive succes and the N
e
/N ratio. Evolution,
59, 1596-1599.
160
HELFMAN, G. S., COLLETTE, B. B. & FACEY, D. E. (1997) The Diversity of Fishes,
Malden, MA, Blackwell Science.
HOARAU, G., BOON, E., JONGMA, D., FERBER, S., PALSSON, J., VAN DER VEER,
H., RIJNSDORP, A., STAM, W. & OLSEN, J. (2005) Low effective population size
and evidence for inbreeding in an overexploited flatfish, plaice, (Pleuronectes
platessa L.). Proceedings of the Royal Society of London B: Biological Sciences,
272, 497-503.
JOHNSON, M. & WERNHAM, J. (1999) Temporal variation of recruits as a basis of
ephemeral genetic heterogeneity in the western rock lobster Panulirus cygnus.
Marine Biology, 135, 133-139.
JOHNSON, M. S. & BLACK, R. (1982) Chaotic genetic patchiness in an intertidal limpet,
Siphonia sp. Marine Biology, 70, 157-164.
JORDE, P. & RYMAN, N. (1996) Demographic genetics of brown trout (Salmo trutta) and
estimation of effective population size from temporal change of allele frequencies.
Genetics, 143, 1369-1381.
JORDE, P. E. & RYMAN, N. (1995) Temporal allele frequency change and estimation of
effective size in populations with overlapping generations. Genetics, 139, 1077-
1090.
KELLOG, J., FLORES, M., HATTOY, B., ROGERS, R., GUSTAFSON, C. & CARLSON
JR., J. (2006) Ocean sport fishing: California marine regulations 2006. California
Department of Fish and Game.
KORDOS, L. & BURTON, R. (1993) Genetic differentiation of Texas Gulf Coast
populations of the blue crab Callinectes sapidus. Marine Biology, 117, 227-233.
LACSON, J. & MORIZOT, D. (1991) Temporal genetic variation in subpopulations of
bicolor damselfish (Stegastes partitus) inhabiting coral reefs in the Florida Keys.
Marine Biology, 110, 353-357.
LANDE, R. (1995) Mutation and conservation. Conservation Biology, 9, 782-791.
LARSON, R. J. & JULIAN, R. M. (1999) Spatial and temporal genetic patchiness in marine
populations and their implications for fisheries management. California Cooperative
Fisheries Investigations, Reports, 40, 94-99.
LEBERG, P. (2005) Genetic approaches for estimating the effective size of populations.
Journal of Wildlife Management, 69, 1385-1399.
LEE, W., CONROY, J., HOWELL, W. & KOCHER, T. (1995) Structure and function of
teleost mitochondrial control regions. Journal of Molecular Evolution, 41, 54-66.
LENFANT, P. & PLANES, S. (2002) Temporal genetic changes between cohorts in a
natural population of a marine fish, Diplodus sargus. Biological Journal of the
Linnean Society, 76, 9-20.
LI, G. & HEDGECOCK, D. (1998) Genetic heterogeneity, detected by PCR-SSCP, among
samples of larval Pacific oysters (Crassostrea gigas) supports the hypothesis of
large variance in reproductive success. Canadian Journal of Fisheries and Aquatic
Science, 55, 1025-1033.
LOVE, M. (1996) Probably More than You Want to Know About the Fishes of the Pacific
Coast, Santa Barbara, CA, Really Big Press.
LOVE, M. S., BROOKS, A., BUSATTO, D., STEPHENS, J. & GREGORY, P. A. (1996)
Aspects of the life histories of the kelp bass, Paralabrax clathratus, and the barred
sand bass, P. nebulifer, from the southern California Bight. Fishery Bulletin, 94,
472-481.
161
LOWE, C. & TOPPING, D. (2003) Movement patterns, home range, and habitat utilization
of adult kelp bass Paralabrax clathratus in a temperate no-take marine reserve.
Marine Ecology Progress Series, 256, 205-216.
LUZIER, C. & WILSON, R. (2004) Analysis of mtDNA haplotypes of kelp bass tests for
sibling-dominated recruitment near marine protected areas of the California Channel
Islands. Marine Ecology Progress Series, 277, 221-230.
MOBERG, P. & BURTON, R. (2000) Genetic heterogeneity among adult and recruit red sea
urchins, Strongylocentrotus franciscanus. Marine Biology, 136, 773-784.
MOSER, H., CHARTER, R., WATSON, W., AMBROSE, D., HILL, K., SMITH, P.,
BUTLER, E., SANDKNOP, M. & CHARTER, S. (2001) The CalCOFI
ichthyoplankton time series: Potential contributions to the management of rocky-
shore fishes. California Cooperative Fisheries Investigations, Reports, 42, 112-128.
NUNNEY, L. (1996) Influence of variation in female fecundity on effective population size.
Biological Journal of the Linnean Society, 59, 411-425.
ODA, D., LAVENBERG, R. & ROUNDS, J. (1993) Reproductive biology of 3 California
species of Paralabrax (Pisces, Serranidae). California Cooperative Oceanic
Fisheries Investigations Reports, 34, 122-132.
OOSTERHOUT, C., HUTCHINSON, W., WILLS, D. & SHIPLEY, P. (2004) Micro-
Checker: software for indentifying and correcting genotyping errors in microsatellite
data. Molecular Ecology Notes, 4, 535-538.
POLLAK, E. (1983) A new method for estimating the effective population size from allele
frequency changes. Genetics, 104, 531-548.
POPPER, A., RAMCHARITAR, J. & CAMPANA, S. (2005) Why otoliths? Insights from
inner ear physiology and fisheries biology. Marine and Freshwater Research, 56,
497-504.
PUJOLAR, J., MAES, G. & VOLCKAERT, F. (2006) Genetic patchiness among recruits in
the European eel Anguilla anguilla. Marine Ecology Progress Series, 307, 209-217.
PURCELL, M., KORNFIELD, I., FOGARTY, M. & PARKER, A. (1996) Interdecadal
heterogeneity in mitochondrial DNA of Atlantic haddock (Melanogrammus
aeglefinus) from Georges Bank. Molecular Marine Biology and Biotechnology, 5,
185-192.
SCHWARTZ, M., TALLMON, D. & LUIKART, G. (1998) Review of DNA-based census
and effective population size estimators. Animal Conservation, 1, 293-299.
SELKOE, K., VOGEL, A. & AFTAB, B. (2005) Eight polymorphic microsatellite markers
for kelp bass, Paralabrax clathratus, amplified in three multiplex polymerase chain
reaction sets. Molecular Ecology Notes, 5, 127-129.
SIVASUNDAR, A., BERMINGHAM, E. & ORTÍ, G. (2001) Population structure and
biogeography of migratory freshwater fishes (Prochilodus: Characiformes) in major
South American rivers. Molecular Ecology, 10, 407-418.
SOULÉ, M. (1980) Thresholds for survival: maintaining fitness and evolutionary potential.
IN SOULÉ, M. & WILCOX, B. (Eds.) Conservaion biology: An evolutionary-
ecological perspective. Sunderland, MA, Sinauer Associates.
SWEARER, S. E., CASELLE, J. E., LEA, D. W. & WARNER, R. R. (1999) Larval
retention and recruitment in an island population of a coral-reef fish. Nature, 402,
799-802.
THOMSON, C. (2001) Human Ecosystem Dimension. IN LEET, W., DEWEES, C.,
KLINGBEIL, R. & LARSON, E. (Eds.) California's living marine resources: A
status report., California Department of Fish and Game.
162
TURNER, T., WARES, J. & GOLD, J. (2002) Genetic effective size is three orders of
magnitude smaller than adult census size in an abundant, estuarine-dependent
marine fish (Sciaenops ocellatus). Genetics, 162, 1329-1339.
VOGEL, A., BURRELL, C., EDMANDS, S. & KIEFER, D. (in prep-a) Correlating
processes associated with ocean triads to the recruitment of the kelp bass,
Paralabrax clathratus.
VOGEL, A. & EDMANDS, S. (in prep-a) The influence of sample size and population
differentiation on statistical power and estimation of F
ST
in marine species.
VOGEL, A. & EDMANDS, S. (in prep-b) Population genetic structure in Paralabrax
clathratus and Clinocottus analis across potential barriers to gene flow.
VOGEL, A., SELKOE, K., ANDERSON, D. & EDMANDS, S. (in prep-b) Development
and inheritance of molecular markers in the kelp bass, Paralabrax clathratus.
WAPLES, R. (1989) A generalized approach for estimating effective population size from
temporal changes in allele frequency. Genetics, 121, 379-391.
WAPLES, R. (1998) Seperating the wheat from the chaff: patterns of genetic differentiation
in high gene flow species. Journal of Heredity, 98, 438-450.
WEIR, B. & COCKERHAM, C. (1984) Estimating F-statistics for the analysis of population
structure. Evolution, 38, 1358-1370.
163
Thesis References
AGOSTINI, V. & BAKUN, A. (2002) 'Ocean Triads' in the Mediterranean Sea: physical
mechanisms potentially structuring reproductive habitat suitability (with example
application to European anchovy, Engraulis encrasicolus. Fisheries Oceanography,
11, 129-142.
AITCHISON, J. (2003) The statistical analysis of compositional data, Caldwell, NJ,
Blackburn Press.
ALLEN, L. & HOVEY, T. (2001) Kelp bass. IN LEET, W., DEWEES, C., KLINGBEIL, R.
& LARSON, E. (Eds.) California's living marine resource: a status report.
California Department of Fish and Game.
ALVARADO-BREMER, J. R. A., MEJUTO, J., GREIG, T. W. & ELY, B. (1996) Global
population structure of the swordfish (Xiphias gladius L.) as revealed by analysis of
the mitochondrial DNA control region. Journal of Experimental Marine Biology and
Ecology, 197, 295-310.
AMMANN, A. (2004) SMURFs: standard monitoring units for the recruitment of temperate
reef fishes. Journal of Experimental Biology and Ecology, 299, 135-154.
ANDERSON, T. (2001) Predator responses, prey refuges, and density-dependent mortality
of a marine fish. Ecology, 82, 245-257.
ANDREWS, A., BURTON, E., KERR, L., CAILLIET, G., COALE, K., LUNDSTROM, C.
& BROWN, T. (2005) Bomb radiocarbon and lead-radium desequilibria in otoliths
of bocaccio rockfish (Sebastes paucispinis): a determination of age and longevity for
a difficult-to-age fish. Marine and Freshwater Research, 56, 517-528.
ANISIMOVA, M., BIELAWSKI, J. & YANG, Z. (2001) Accuracy and power of the
likelihood ratio test in detecting adaptive molecular evolution. Molecular Biology
and Evolution, 18, 1585-1592.
ANISIMOVA, M., BIELAWSKI, J. & YANG, Z. (2002) Accuracy and power of Bayes
prediction of amino acid sites under positive selection. Molecular Biology and
Evolution, 19, 950-958.
AVISE, J. (2000) Phylogeography: the history and formation of species, Cambridge, MA,
Harvard University Press.
AYERS, K. & WATER, J. (2005) Marine biogeographic disjunction in central New Zealand.
Marine Biology, 147, 1045-1052.
BAKUN, A. (1996a) Patterns in the Ocean, Ocean Processes and Marine Population
Dynamics, La Paz, BCS, Mexico, California Sea Grant/CIB.
BAKUN, A. (1996b) Patterns in the ocean: Ocean processes and marine population
dynamics, La Paz, BCS, Mexico, California Sea Grant/CIB.
BAKUN, A. & NELSON, C. (1977) Climatology of upwelling related processes off Baja
California. California Cooperative Fisheries Investigations, Reports, 19, 107-127.
BARTLEY, D., BAGLEY, M., GALL, G. & BENTLEY, B. (1992) Use of linkage
disequilibrium data to estimate effective size of hatchery and natural fish
populations. Conservation Biology, 6, 365-375.
BAUMS, I., HUGHES, C. & HELLBERG, M. (2005) Mendelian microsatellite loci for the
Caribbean coral Acropora palmata. Marine Ecology Progress Series, 288, 115-127.
BAZIN, E., GLÉMIN, S. & GALTIER, N. (2006) Population size does not influence
mitochondrial genetic diversity in animals. Science, 312, 570-572.
164
BECKWITT, R. (1983) Genetic Structure of Genyonemus lineatus, Seriphus politus
(Sciaenidae) and Paralabrax clathratus (Serranidae) in Southern California. Copeia,
3, 691-696.
BEKKEVOLD, D., HANSEN, M. & LOESCHCKE, V. (2002) Male reproductive
competition in spawning aggregations of cod (Gadus morhua, L.). Molecular
Ecology, 11, 91-102.
BELKHIR, K., BORSA, P., CHIKHI, L., RAUFASTE, N. & BONHOMME, F. (2004)
Genetix 4.05, logiciel sous Windows TM pour la génétique des populations. IN
LABORATOIRE GÉNOME, P., INTERACTIONS, CNRS UMR 5171 (Ed.)
Montpellier (France), Université de Montpellier II.
BENTZEN, P. (1998) Seeking evidence of local stock structure using molecular genetic
methods. IN VON HERBING, I., KORNFIELD, I., TUPPER, M. & WILSON, J.
(Eds.) The implications of localized fishery stocks. Portland, Maine, Natural
resource, agriculture, and engineering service.
BERKELEY, S., CHAPMAN, C. & SOGARD, S. (2004) Maternal age as a determinant of
larval growth and survival in a marine fish, Sebastes melanops. Ecology, 85, 1258-
1264.
BERNARDI, G. (2000) Barriers to gene flow in Embiotoca jacksoni, a marine fish lacking a
pelagic larval stage. Evolution, 54, 226-237.
BERNARDI, G., FINDLEY, L. & ROCHA-OLIVARES, A. (2003) Vicariance and dispersal
across Baja California in disjunct marine fish populations. Evolution, 57, 1599-1609.
BERNARDI, G. & TALLEY, D. (2000) Genetic evidence for limited dispersal in the coastal
California killifish, Fundulus parvipinnis. Journal of Experimental Marine Biology
and Ecology, 255, 187-199.
BOWEN, B. W. & AVISE, J. C. (1990) Genetic structure of Atlantic and Gulf of Mexico
populations of sea bass, menhaden, and sturgeon: influences of zoogeographic
factors and life-history patterns. Marine Biology, 107, 371-381.
BREIMAN, L. & FRIEDMAN, J. (1985) Estimating optimal transformations for multiple
regression and correlation. Journal of the American Statistical Association, 80, 580-
598.
BUONACCORSI, V., STARKEY, E. & GRAVES, J. (2001) Mitochondrial and nuclear
DNA analysis of population subdivision among young-of-the-year Spanish mackerel
(Scomberomorus maculatus) from the western Atlantic and Gulf of Mexico. Marine
Biology, 138, 37-45.
BURTON, R. (1996) Molecular tools in marine ecology. Journal of Experimental Marine
Biology and Ecology, 200, 85-101.
BURTON, R. (1998) Intraspecific phylogeography across the Point Conception
biogeographic boundary. Evolution, 52, 734-745.
CAI, J., SMITH, D., XIA, X. & YUEN, K. (2005) MBEToobox: a Matlab toolbox for
sequence data analysis in molecular biology and evolution. BMC Bioinformatics, 6,
doi:10.1186/1471-2105-6-64.
CALDEIRA, R., MARCHESIELLO, P., NEZLIN, N., DIGIACOMO, P. & MCWILLIAMS,
J. (2005) Island wakes in the Southern California Bight. Journal of Geophysical
Research-Oceans, 110, doi:10.1029/2004JC002675.
CARR, M. (1994) Effects of macroalgal dynamics on recruitment of a temperate reef fish.
Ecology, 75, 1320-1333.
165
CERMEÑO, P., URIARTE, A., DE MURGUÍA, A. & MORALES-NIN, B. (2003)
Validation of daily increment formation in otoliths of juvenile and adult European
anchovy. Journal of Fish Biology, 62, 679-691.
CHAKRABORTY, R. & DANKER-HOPFE, H. (1991) Analysis of population structure: a
comparative analysis of different estimators of Wright's fixation indices. IN RAO,
C. & CHAKRABORTY, R. (Eds.) Handbook of statistics 8: Statistical methods in
biological and medical sciences. Amsterdam, the Netherlands, Elsevier Science
Publishers B.V.
CHAPMAN, R. W. (1989) Spatial and temporal variation of mitochondrial DNA haplotype
frequences in the striped bass (Morone saxatilis) 1982 year class. Copeia, 2, 344-
348.
COCKERHAM, C. (1969) Variance of gene frequencies. Evolution, 23, 72-84.
COCKERHAM, C. & WEIR, B. (1986) Estimation of inbreeding parameters in stratified
populations. Annals of Human Genetics, 50, 271-281.
COCKERHAM, C. & WEIR, B. (1993) Estimation of gene flow from F-statistics. Evolution,
47, 855-863.
COHEN, J. (1988) Statistical power analysis for the behavioral sciences, Mahwah, New
Jersey, Lawrence Erlbaum Associates.
COLE, J. (1999) Environmental conditions, satellite imagery, and clupeoid recruitment in
the northern Benguela upwelling system. Fisheries Oceanography, 8, 25-38.
COLE, J. & MCGLADE, J. (1998a) Clupeoid population variability, the environment and
satellite imagery in coastal upwelling systems. Reviews in Fish Biology and
Fisheries, 8, 445-471.
COLE, J. & MCGLADE, J. (1998b) Temporal and spatial patterning of sea surface
temperature in the northern Benguela upwelling system: possible environmental
indicators of clupeoid production. South African Journal of Marine Science, 19,
143-157.
COPE, J. (2004) Population genetics and phylogeography of the blue rockfish (Sebastes
mystinus) from Washington to California. Canadian Journal of Fisheries and
Aquatic Science, 61, 332-342.
CORDES, J. F. & ALLEN, L. G. (1997) Estimates of age, growth, and settlement from
otoliths of young-of-the-year kelp bass (Paralabrax clathratus). Bulletin of the
Southern California Academy of Sciences, 96, 43-60.
COX, D. & STUART, A. (1955) Some quick tests for trend and dispersion. Biometrika, 42,
80-95.
CROW, J. & KIMURA, M. (1970) An introduction to population genetics theory,
Minneapolis, Burgess.
CURY, P. & ROY, C. (1989) Optimal environmental window and pelagic fish recruitment
success in upwell areas. Canadian Journal of Fisheries and Aquatic Science, 46,
670-680.
CURY, P., ROY, C., MENDESSOHN, R., BAKUN, A., HUSBY, D. & PARRISH, R.
(1995) Moderate is better: exploring nonlinear climatic effects on the Californian
northern anchovy (Engraulis mordax). Canadian Journal of Fisheries and Aquatic
Science, 121, 417-424.
DAVID, P., PERDIEU, M., PERNOT, A. & JARNE, P. (1997) Fine-grained spatial and
temporal population genetic structure in the marine bivalve Spisula ovalis.
Evolution, 51, 1318-1322.
166
DAVIS, G., KUSHNER, D., MONDRAGON, J., MONDRAGON, J., LERMA, D. &
RICHARDS, D. (1997) Kelp forest monitoring handbook volume 1: sampling
protocol. Ventura, California, National Park Service, Channel Islands National Park.
DAVIS, J. & LEVIN, L. (2002) Importance of pre-recruitment life-history stages to
population dynamics of the woolly sculpin, Clinocottus analis. Marine Ecology
Progress Series, 234, 229-246.
DE MEEUS, T., HUMAIR, P., GRUNAU, C., DELAYE, C. & RENAUD, F. (2004) Non-
Mendelian transmission of alleles at microsatellite loci: an example in Ixodes
ricinus, the vector of Lyme disease. International Journal of Parisitology, 34, 943-
950.
DEL MONTE LUNA, P. (2004) Caracterización del centro de actividad biológica del Golfo
de Ulloa, B.C.S., bajo un enfoque del modelación ecológica. Instituto Politécnico
Nacional. La Paz, B.C.S., CICIMAR.
DEWOODY, J. & AVISE, J. (2000) Microsatellite variation in marine, freshwater and
anadromous fishes compared with other animals. Journal of Fish Biology, 56, 461-
473.
DI LORENZO, E. (2003) Seasonal dynamics of the surface circulation in the Southern
California Current system. Deep-Sea Research II, 50, 2371-2388.
DIZON, A., TAYLOR, B. & O'CORRY-CROWE, G. (1995) Why statistical power is
necessary to link analyses of molecular variation to decisions about population
structure. American Fisheries Society Symposium, 17, 288-294.
DURAZO, R. & BAUMGARTNER, T. (2002) Evolution of oceanographic conditions off
Baja California: 1997-1999. Progress in Oceanography, 54, 7-31.
EDDY, S. (2005) A model of the statistical power of comparative genome sequence
analysis. Public Library of Science, Biology, 3, 95-102.
EDMANDS, S., MOBERG, P. & BURTON, R. (1996) Allozyme and mitochondrial DNA
evidence of population subdivision in the purple sea urchin Strongylocentrotus
purpuratus. Marine Biology, 126, 443-450.
EPPLEY, R. (1992) Photosynthesis and new production in the Southern California Bight.
Progress in Oceanography, 30, 117-150.
ERISMAN, B. & ALLEN, L. (2006) Reproductive behaviour of a temperate serranid fish,
Paralabrax clathratus (Girard), from Santa Catalina Island, California, USA.
Journal of Fish Biology, 68, 157-184.
ESCHMEYER, W. N., HERALD, E. S. & HAMMANN, H. (1983) A Field Guide to Pacific
Coast Fishes, New York, New York, Houghton Mifflin Company.
ESPINOSA-CARREON, T., STRUB, P., BEIER, E., OCAMPO-TORRES, F. &
GAXIOLA-CASTRO, G. (2004) Seasonal and interannual variability of satellite-
derived chlorophyll piment, surface height, and temperature off Baja California.
Journal of Geophysical Research-Oceans, 109, C03039,
doi:10.1029/2003JC002105.
EWENS, W. (1972) The sampling theory of selectively neutral alleles. Theoretical
Population Biology, 3, 87-112.
EXCOFFIER, L., LAVAL, L. & SCHNEIDER, S. (2005) Arlequin ver. 3.0: An integrated
software package for population genetics data analysis. Evolutionary Bioinformatics
Online, 1, 47-50.
EXCOFFIER, L., SMOUSE, P. & QUATTRO, J. (1992) Analysis of molecular variance
inferred from metric distances among DNA haplotypes: application to human
mitochondrial DNA restriction data. Genetics, 131, 479-491.
167
FAUVELOT, C. & PLANES, S. (2002) Understanding origins of present-day genetic
structure in marine fish: biologically or historically driven patterns? Marine Biology,
141, 773-788.
FELDHEIM, K., GRUBER, S. & ASHLEY, M. (2001) Population genetic structure of the
lemon shark (Negaprion brevirostris) in the western Atlantic: DNA microsatellite
variation. Molecular Ecology, 10, 295-303.
FINDLAY, A. & ALLEN, L. G. (2002) Temporal patterns of settlement in the temperate
reef fish Paralabrax clathratus. Marine Ecology Progress Series, 238, 237-248.
FORRESTER, G. & SWEARER, S. E. (2002) Trace elements in otoliths indicate the use of
open-coast versus bay nursery habitats by juvenile California halibut. Marine
Ecology Progress Series, 241, 201-213.
FRANKLIN, I. (1980) Evolutionary changes in small populations. IN SOULÉ, M. &
WILCOX, B. (Eds.) Conservaion biology: An evolutionary-ecological perspective.
Sunderland, MA, Sinauer Associates.
GERARD, P., SMITH, D. & WEERAKKODY, G. (1998) Limits of retrospective power
analysis. Journal of Wildlife Management, 62, 801-807.
GILBERT-HORVATH, E., LARSON, R. & GARZA, J. (2006) Temporal recruitment
patterns and gene flow in kelp rockfish (Sebastes atrovirens). Molecular Ecology,
doi: 10.1111/j.1365-294X.2006.03033.x.
GILLANDERS, B. (2002) Temporal and spatial variability in elemental composition of
otoliths: implications for determining stock identity and connectivity of populations.
Canadian Journal of Fisheries and Aquatic Science, 59, 669-679.
GOODNIGHT, K. & QUELLER, D. (1999) Computer software for performing likelihood
tests of pedigree relationship using genetic markers. Molecular Ecology, 8, 1231-
1234.
GOUDET, J., RAYMOND, M., DE MEEÜS, T. & ROUSSET, F. (1996) Testing
differentiation in diploid populations. Genetics, 144, 1933-1940.
GOULD, A. (2001) Sample size re-estimation: recent developments and practical
considerations. Statistics in Medicine, 20, 2625-2643.
GRANT, W. & BOWEN, B. (1998) Shallow population histories in deep evolutionary
lineages of marine fishes: insights from sardines and anchovies and lessons for
conservation. The Journal of Heredity, 89, 415-426.
GRAVES, J. & MCDOWELL, J. (2003) Stock structure of the world's istiophorid billfishes:
a genetic perspective. Marine and Freshwater Research, 54, 287-298.
GROSSMAN, J. & MACKENZIE, F. (2005) The randomized controlled trial. Perspectives
in Biology and Medicine, 48, 516-534.
GROTHUES, T. M. (1994) An investigation into the phenetic structure and larval dispersal
patterns of the kelp bass, Paralabrax clathratus. Marine Biology. Los Angeles,
California State University, Northridge.
HARMS, S. & WINANT, C. (1998) Characteristic patterns of the circulation in the Santa
Barbara Channel. Journal of Geophysical Research-Oceans, 103, 3041-3065.
HARTL, D. & CLARK, A. (1989) Principals of Population Genetics, Sunderland, Sinauer
Associates, Inc.
HAUSER, L., ADCOCK, G., SMITH, P., BERNAL RAMÍREZ, J. & CARVALHO, G.
(2002) Loss of microsatellite diversity and low effective population size in an
overexploited population of New Zealand snapper (Pagrus auratus). Proceedings of
the National Academy of Sciences of the United States of America, 99.
168
HAYES, J. & STEIDL, R. (1997) Statistical power analysis and amphibian population
trends. Conservation Biology, 11, 273-275.
HEDGECOCK, D. (1986) Is gene flow from pelagic larval dispersal important in the
adaptation and evolution of marine invertebrates? Bulletin of Marine Science, 39,
550-564.
HEDGECOCK, D. (1994a) Does variance in reproductive success limit effective population
sizes of marine organisms? IN BEAUMONT, A. (Ed.) Genetics and Evolution of
Aquatic Organisms. London, Chapman and Hall.
HEDGECOCK, D. (1994b) Temporal and spatial genetic structure of marine animal
populations in the California current. California Cooperative Fisheries
Investigations, Reports, 35, 73-81.
HEDGECOCK, D., CHOW, V. & WAPLES, R. (1992) Effective population numbers of
shellfish broodstocks estimated from temporal variance in allelic frequencies.
Aquaculture, 108, 215-232.
HEDGECOCK, D., HUTCHINSON, E., LI, G., SLY, F. & NELSON, K. (1994) The central
stock of northern anchovy Engraulis mordax is not a randomly mating population.
California Cooperative Fisheries Investigations, Reports, 35, 121-136.
HEDGECOCK, D., LAUNEY, S., PUDOVKIN, A., NACIRI, Y., LAPÈGUE, S. &
BONHOMME, F. (2006) Small effective number of parents (N
b
) inferred for a
naturally spawned cohort of juvenile European flat oysters Ostrea edulis. Marine
Biology, doi: 10.1007/s00227-006-0441-y.
HEDGECOCK, D. & SLY, F. (1990) Genetic drift and effective population sizes of
hatchery-propagated stocks of the Pacific oyster, Crassostrea gigas. Aquaculture,
88, 21-38.
HEDRICK, P. (1999) Perspective: Highly variable loci and their interpretation in evolution
and conservation. Evolution, 53, 313-318.
HEDRICK, P. (2000) Genetics of populations, London, Jones and Bartlett Publishers
International.
HEDRICK, P. (2005a) Large variance in reproductive succes and the N
e
/N ratio. Evolution,
59, 1596-1599.
HEDRICK, P. (2005b) A standardized genetic differentiation measure. Evolution, 59, 1633-
1638.
HELFMAN, G. S., COLLETTE, B. B. & FACEY, D. E. (1997) The Diversity of Fishes,
Malden, MA, Blackwell Science.
HICKEY, B., DOBBINS, E. & ALLEN, S. (2003) Local and remote forcing of currents and
temperature in the central Southern California Bight. Journal of Geophysical
Research-Oceans, 108, doi:10.1029/200JC000313.
HOARAU, G., BOON, E., JONGMA, D., FERBER, S., PALSSON, J., VAN DER VEER,
H., RIJNSDORP, A., STAM, W. & OLSEN, J. (2005) Low effective population size
and evidence for inbreeding in an overexploited flatfish, plaice, (Pleuronectes
platessa L.). Proceedings of the Royal Society of London B: Biological Sciences,
272, 497-503.
HOENIG, J. & HEISEY, D. (2001) The abuse of power: the pervasive fallacy of power
calculations for data analysis. The American Statistician, 55, 1-6.
HUANG, D. & BERNARDI, G. (2001) Disjunct Sea of Cortez-Pacific Ocean Gillychthys
mirabilis populations and the evolutionary origin of the their Sea of Cortez endemic
relative. Marine Biology, 138.
169
HUBBS, C. (1926) Descriptions of new genera of cottoid fishes related to Artedius.
Occasional Papers of the Museum of Zoology, University of Michigan, 1-16.
HUELSENBECK, J. & RONQUIST, F. (2001) MRBAYES: Bayesian inference of
phylogeny. Bioinformatics, 17, 754-755.
JENNIONS, M. & MØLLER, A. (2002a) Publication bias in ecology and evolution: an
empirical assessment using the "trim and fill" method. Biological Reviews, 77, 211-
222.
JENNIONS, M. & MØLLER, A. (2002b) Relationships fade with time: a meta-analysis of
temporal trends in publication in ecology and evolution. Proceedings of the Royal
Society of London B: Biological Sciences, 269, 43-48.
JENNIONS, M. & MØLLER, A. (2003) A survey of the statistical power of research in
behavioral ecology and animal behavior. Behavioral Ecology, 14, 438-445.
JOHNSON, M. & WERNHAM, J. (1999) Temporal variation of recruits as a basis of
ephemeral genetic heterogeneity in the western rock lobster Panulirus cygnus.
Marine Biology, 135, 133-139.
JOHNSON, M. S. & BLACK, R. (1982) Chaotic genetic patchiness in an intertidal limpet,
Siphonia sp. Marine Biology, 70, 157-164.
JORDE, P. & RYMAN, N. (1996) Demographic genetics of brown trout (Salmo trutta) and
estimation of effective population size from temporal change of allele frequencies.
Genetics, 143, 1369-1381.
JORDE, P. E. & RYMAN, N. (1995) Temporal allele frequency change and estimation of
effective size in populations with overlapping generations. Genetics, 139, 1077-
1090.
JUKES, T. & CANTOR, C. (1969) Evolution of protein molecules. IN MUNRO, H. (Ed.)
Mammalian Protein Metabolism. New York, Academic Press.
KELLOG, J., FLORES, M., HATTOY, B., ROGERS, R., GUSTAFSON, C. & CARLSON
JR., J. (2006) Ocean sport fishing: California marine regulations 2006. California
Department of Fish and Game.
KELLY, N. & KIRA, G. (1998) The Baja catch: A fishing, travel and remote camping
manual for Baja California, Valley Center, CA, Apples and Oranges, Inc.
KIMURA, M. (1980) A simple method for estimating evolutionary rate of of base
substitution through comparative studies of nucleotide sequences. Journal of
Molecular Evolution, 16, 111-120.
KNUTSEN, H., JORDE, P., ANDRE, C. & STENSETH, N. (2003) Fine-scaled
geographical population structuring in a highly mobile marine species: the Atlantic
cod. Molecular Ecology, 12, 385-394.
KORDOS, L. & BURTON, R. (1993) Genetic differentiation of Texas Gulf Coast
populations of the blue crab Callinectes sapidus. Marine Biology, 117, 227-233.
KUHNER, M., YAMATO, J. & FELSENSTEIN, J. (1998) Maximum likelihood estimation
of population growth rates based on the coalescent. Genetics, 149, 429-434.
LACSON, J. & MORIZOT, D. (1991) Temporal genetic variation in subpopulations of
bicolor damselfish (Stegastes partitus) inhabiting coral reefs in the Florida Keys.
Marine Biology, 110, 353-357.
LANDE, R. (1995) Mutation and conservation. Conservation Biology, 9, 782-791.
LARSON, R. J. & JULIAN, R. M. (1999) Spatial and temporal genetic patchiness in marine
populations and their implications for fisheries management. California Cooperative
Fisheries Investigations, Reports, 40, 94-99.
170
LAURENT, V., VOISIN, M. & PLANES, S. (2006) Genetic clines in the Bay of Biscay
provide estimates of migration for Sardina pilchardus. Journal of Heredity, 97, 81-
88.
LEBERG, P. (2005) Genetic approaches for estimating the effective size of populations.
Journal of Wildlife Management, 69, 1385-1399.
LEE, W., CONROY, J., HOWELL, W. & KOCHER, T. (1995) Structure and function of
teleost mitochondrial control regions. Journal of Molecular Evolution, 41, 54-66.
LEEBENS-MACK, J. & DEPAMPHILIS, C. (2002) Power analysis of tests for loss of
selective constraint in cave crayfish and nonphotosynthetic plant lineages.
Molecular Biology and Evolution, 19, 1292-1302.
LEGAARD, K. & THOMAS, A. (2006) Spatial patterns in seasonal and interannual
variability of chlorophyll and sea surface temperature in the California Current.
Journal of Geophysical Research-Oceans, 111, doi:10.1029/2005JC003282.
LENFANT, P. & PLANES, S. (2002) Temporal genetic changes between cohorts in a
natural population of a marine fish, Diplodus sargus. Biological Journal of the
Linnean Society, 76, 9-20.
LI, G. & HEDGECOCK, D. (1998) Genetic heterogeneity, detected by PCR-SSCP, among
samples of larval Pacific oysters (Crassostrea gigas) supports the hypothesis of
large variance in reproductive success. Canadian Journal of Fisheries and Aquatic
Science, 55, 1025-1033.
LLUCH-COTA, S., LLUCH-COTA, D., LLUCH-BELDA, D., NEVAREZ-MARTINEZ,
M., PARES-SIERRA, A. & HERNANDEZ-VAZQUEZ, S. (1999) Variability of
sardine catch as related to enrichment, concentration and retention processes in the
central Gulf of California. California Cooperative Fisheries Investigations, Reports,
40, 184-190.
LOVE, M. (1996) Probably More than You Want to Know About the Fishes of the Pacific
Coast, Santa Barbara, CA, Really Big Press.
LOVE, M., NISHIMOTO, M., SCHROEDER, D. & CASELLE, J. (1999) The ecological
role of natural reefs and oil and gas production platforms on rocky reef fishes in
southern California: Final Interim Report., U.S. Department of the Interior, U.S.
Geological Survey, Biological Resources Division.
LOVE, M. S., BROOKS, A., BUSATTO, D., STEPHENS, J. & GREGORY, P. A. (1996)
Aspects of the life histories of the kelp bass, Paralabrax clathratus, and the barred
sand bass, P. nebulifer, from the southern California Bight. Fishery Bulletin, 94,
472-481.
LOWE, C. & TOPPING, D. (2003) Movement patterns, home range, and habitat utilization
of adult kelp bass Paralabrax clathratus in a temperate no-take marine reserve.
Marine Ecology Progress Series, 256, 205-216.
LOWE, C., TOPPING, D., CARTAMIL, D. & PAPASTAMATIOU, Y. (2003) Movement
patterns, home range, and habitat utilization of adult kelp bass Paralabrax clathratus
in a temperate no-take marine reserve. Marine Ecology Progress Series, 256, 205-
216.
LUZIER, C. & WILSON, R. (2004) Analysis of mtDNA haplotypes of kelp bass tests for
sibling-dominated recruitment near marine protected areas of the California Channel
Islands. Marine Ecology Progress Series, 277, 221-230.
MACKENZIE, B. (2000) Turbulence, larval fish ecology and fisheries recruitment: a review
of field studies. Oceanologica Acta, 23, 357-375.
171
MARSHALL, T., SLATE, J., LEB, K. & PEMBERTON, J. (1998) Statistical confidence for
likelihood-based paternity inference in natural populations. Molecular Ecology, 7,
639-655.
MCPHERSON, A., STEPHENSON, R. & TAGGART, C. (2003) Genetically different
Atlantic herring Clupea harengus spawning waves. Marine Ecology Progress
Series, 247, 303-309.
MENDESSOHN, R., SCHWING, F. & BOGRAD, S. (2003) Spatial structure of subsurface
temperature variability in the California Current, 1950-1993. Journal of Geophysical
Research-Oceans, 108, 3093, doi:10.1029/2002JC001568.
MILLER, A., MCWILLIAMS, J., SCHNEIDER, N., ALLEN, J., BARTH, J.,
BEARDSLEY, R., CHAVEZ, F., CHERESKIN, T., EDWARDS, C., HANEY, R.,
KELLY, K., KINDLE, J., LY, L., MOISAN, J., NOBLE, M., NIILER, P., OEY, L.,
SCHWING, F., SHEARMAN, R. & SWENSON, M. (1999) Observing and
modeling the California Current System. EOS, Transactions, American Geophysical
Union, 80, 533-539.
MOBERG, P. & BURTON, R. (2000) Genetic heterogeneity among adult and recruit red sea
urchins, Strongylocentrotus franciscanus. Marine Biology, 136, 773-784.
MORGAN, M. (1997) Genetic variation among nine populations of California halibut,
Paralichthys californicus. Biology. Northridge, CA, California State University,
Northridge.
MORIN, P., LUIKART, G., WAYNE, R. & GROUP, T. S. W. (2004) SNPs in ecology,
evolution and conservation. Trends in Ecology and Evolution, 19, 208-216.
MOSER, H., CHARTER, R., WATSON, W., AMBROSE, D., HILL, K., SMITH, P.,
BUTLER, E., SANDKNOP, M. & CHARTER, S. (2001) The CalCOFI
ichthyoplankton time series: Potential contributions to the management of rocky-
shore fishes. California Cooperative Fisheries Investigations, Reports, 42, 112-128.
MYERS, R. (1998) When do environmental-recruitment correlations work? Reviews in Fish
Biology and Fisheries, 8, 285-305.
NAKAGAWA, S. & FOSTER, T. (2004) The case against retrospective statistical power
analyses with an introduction to power analysis. Acta Ethologica, 7, 103-108.
NEFF, B., REPKA, J. & GROSS, M. (2000) Statistical confidence in parentage analysis
with incomplete sampling: how many loci and offspring are needed? Molecular
Ecology, 9, 529-539.
NEI, M. (1977) F-statistics and analysis of gene diversity in subdivided populations. Annals
of Human Genetics, 11, 307-325.
NEI, M. (1987) Molecular Evolutionary Genetics, New York, Columbia University Press.
NEI, M. & CHESSER, R. (1983) Estimation of fixation indices and gene diversities. Annals
of Human Genetics, 47, 253-259.
NEIGEL, J. (2002) Is F
ST
obsolete? Conservation Genetics, 3, 167-173.
NEWMAN, S., STECKIS, R., EDMONDS, J. & LLOYD, J. (2000) Stock structure of the
goldband snapper Pristipomoides multidens (Pisces: Lutjanidae) from the waters of
northern and western Australia by stable isotope ratio analysis of sagittal otolith
carbonate. Marine Ecology Progress Series, 198.
NEZLIN, N. & MCWILLIAMS, J. (2003) Satellite data, Empirical Orthogonal Functions,
and the 1997-1998 El Niño off California. Remote Sensing of Environment, 84, 234-
254.
172
NISHIMOTO, M. & WASHBURN, L. (2002) Patterns of coastal eddy circulation and
abundace of pelagic juvenile fish in the Santa Barbara Channel, California, USA.
Marine Ecology Progress Series, 241, 183-199.
NUNNEY, L. (1996) Influence of variation in female fecundity on effective population size.
Biological Journal of the Linnean Society, 59, 411-425.
NURBAEV, S. & BALANOVSKAIA, E. (1998) Interpopulation diversity of the gene pool:
beta distribution of Wright's F
ST
statistics. Russian Journal of Human Genetics, 34,
1004-1008.
ODA, D., LAVENBERG, R. & ROUNDS, J. (1993) Reproductive biology of 3 California
species of Paralabrax (Pisces, Serranidae). California Cooperative Oceanic
Fisheries Investigations Reports, 34, 122-132.
OOSTERHOUT, C., HUTCHINSON, W., WILLS, D. & SHIPLEY, P. (2004) Micro-
Checker: software for indentifying and correcting genotyping errors in microsatellite
data. Molecular Ecology Notes, 4, 535-538.
PAETKAU, D., SLADE, R., BURDEN, M. & ESTOUP, A. (2004) Genetic assignment
methods for the direct, real-time estimation of migration rate: a simulation-based
exploration of accuracy and power. Molecular Ecology, 13, 55-65.
PALUMBI, S. (2003) Population genetics, demographic connectivity, and the design of
marine reserves. Ecological Applications, 13, S146-S158.
PARRISH, R., NELSON, C. & BAKUN, A. (1981) Transport mechanisms and reproductive
success of fishes in the California Current. Biological Oceanography, 1, 175-203.
PEREZ-ENRIQUEZ, R., VEGA, A., AVILA, S. & SANDOVAL, J. (2001) Population
genetics of red spiny lobster (Panulirus interruptus) along the Baja California
peninsula, Mexico. Marine and Freshwater Research, 52, 1541-1549.
PETERSON, W. & SCHWING, F. (2003) A new climate regime in northeast Pacific
systems. Geophysical Research Letters, 30, doi:10.1029/2003GL017528.
PLANES, S. & LENFANT, P. (2002) Temporal change in the genetic structure between and
within cohorts of a marine fish, Diplodus sargus, induced by a large variance in
individual reproductive success. Molecular Ecology, 11, 1515-1524.
POLLAK, E. (1983) A new method for estimating the effective population size from allele
frequency changes. Genetics, 104, 531-548.
POPPER, A., RAMCHARITAR, J. & CAMPANA, S. (2005) Why otoliths? Insights from
inner ear physiology and fisheries biology. Marine and Freshwater Research, 56,
497-504.
PUJOLAR, J., MAES, G. & VOLCKAERT, F. (2006) Genetic patchiness among recruits in
the European eel Anguilla anguilla. Marine Ecology Progress Series, 307, 209-217.
PURCELL, M., KORNFIELD, I., FOGARTY, M. & PARKER, A. (1996) Interdecadal
heterogeneity in mitochondrial DNA of Atlantic haddock (Melanogrammus
aeglefinus) from Georges Bank. Molecular Marine Biology and Biotechnology, 5,
185-192.
RAUFASTE, N. & BONHOMME, F. (2000) Properties of bias and variance of two
multiallelic estimators of F
ST
. Theoretical Population Biology, 57, 285-296.
REED, J., MILLS, L., DUNNING JR., J., MENGES, E., MCKELVEY, K., FRYE, R.,
BEISSINGER, S., ANSTETT, M. & MILLER, P. (2002) Emerging issues in
population viability analysis. Conservation Biology, 16, 7-19.
REYNOLDS, J., WEIR, B. & COCKERHAM, C. (1983) Estimation of the coancestry
coefficient: basis for a short-term genetic distance. Genetics, 105, 767-779.
173
ROBERTSON, A. & HILL, W. (1984) Deviations from Hardy-Weinberg proportions:
sampling variances and use in estimation of inbreeding coefficients. Genetics, 107,
703-718.
ROCHA-OLIVARES, A., ROSENBLATTT, R. & VETTER, R. (1999) Cryptic species of
rockfish (Sebastes, Scorpaenidae) in the southern hemisphere inferred from
mitochondrial lineages. Journal of Heredity, 90, 404-411.
ROCHA-OLIVARES, A. & VETTER, R. D. (1999) Effects of oceanographic circulation on
the gene flow, genetic structure, and phyogeography of the rosethorn rockfish
(Sebastes helvomaculatus). Canadian Journal of Fisheries and Aquatic Science, 56,
803-813.
ROY, C., CURY, P. & KIFANI, S. (1992) Pelagic fish recruitment success and reproductive
strategy in upwelling areas: environmental comprimises. South African Journal of
Marine Science, 12, 135-146.
RYMAN, N., PALM, S., ANDRÉ, C., CARVALHO, G., DAHLGREN, T., JORDE, P.,
LAIKRE, L., LARSSON, L., PALMÉ, A. & RUZZANTE, D. (2006) Power for
detecting genetic divergence: differences between statistical methods and marker
loci. Molecular Ecology, 15, 2031-2045.
SALOMON, M. (2005) Gene flow among populations of the spotted sand bass, Paralabrax
maculatofasciatus, from southern California and two populations along the pacific
coast of Baja, California. Biology. Northridge, CA, California State University,
Northridge.
SCHREY, A. & HEIST, E. (2003) Microsatellite analysis of population structure in the
shortfin mako (Isurus oxyrinchus). Canadian Journal of Fisheries and Aquatic
Science, 60, 670-675.
SCHUELKE, M. (2000) An economic mathod for the fluorescent labeling of PCR
fragments. Nature Biotechnology, 18, 233-234.
SCHWARTZ, M., TALLMON, D. & LUIKART, G. (1998) Review of DNA-based census
and effective population size estimators. Animal Conservation, 1, 293-299.
SELKOE, K. & TOONEN, R. (2006) Microsatellites for ecologists: a practical guide to
using and evaluating microsatellite markers. Ecology Letters, doi: 10.111/j.1461-
0248.2006.00889.x.
SELKOE, K., VOGEL, A. & AFTAB, B. (2005) Eight polymorphic microsatellite markers
for kelp bass, Paralabrax clathratus, amplified in three multiplex polymerase chain
reaction sets. Molecular Ecology Notes, 5, 127-129.
SHAW, P., ARKHIPKIN, A. & AL-KHAIRULLA, H. (2004) Genetic structuring of
Patagonian toothfish populations in the Southwest Atlantic Ocean: the effect of the
Antarctic Polar Front and deep-water troughs as barriers to genetic exchange.
Molecular Ecology, 13, 3293-3303.
SIRABELLA, P., GUILIANI, A., COLOSIMO, A. & DIPPNER, J. (2001) Breaking down
the climate effects on cod recruitment by principal component analysis and
canonical correlation. Marine Ecology Progress Series, 216, 213-222.
SIVASUNDAR, A., BERMINGHAM, E. & ORTÍ, G. (2001) Population structure and
biogeography of migratory freshwater fishes (Prochilodus: Characiformes) in major
South American rivers. Molecular Ecology, 10, 407-418.
SLATKIN, M. (1987) Gene flow and the geographic structure of natural populations.
Science, 236, 787-792.
SLATKIN, M. (1995) A measure of population subdivision based on microsatellite allele
frequencies. Genetics, 139, 457-462.
174
SLATKIN, M. (1996) A correction to the exact test based on the Ewens sampling
distribution. Genetical Research, 68, 259-260.
SMITH, K., ALBERTS, S., BAYES, M., BRUFORD, M., ALTMANN, J. & OBER, C.
(2000) Cross-species amplification, non-invasive genotyping, and non-Mendelian
inheritance of human STRPs in Savannah baboons. American Journal of
Primatology, 51, 219-227.
SOULÉ, M. (1980) Thresholds for survival: maintaining fitness and evolutionary potential.
IN SOULÉ, M. & WILCOX, B. (Eds.) Conservaion biology: An evolutionary-
ecological perspective. Sunderland, MA, Sinauer Associates.
STAR, B., APTE, S. & GARDNER, J. (2003) Genetic structuring among populations of the
greenshell mussel Perna canaliculus revealed by analysis of randomly amplified
polymorphic DNA. Marine Ecology Progress Series, 249, 171-182.
STEELE, M., MALONE, J., FINDLAY, A., CARR, M. & FORRESTER, G. (2002) A
simple method for estimating larval supply in reef fishes and a preliminary test of
population limitation by larval delivery in the kelp bass, Paralabrax clathratus.
Marine Ecology Progress Series, 235, 195-203.
STEPIEN, C., ROSENBLATTT, R. & BARGMEYER, B. (2001) Phylogeography of the
spotted sand bass, Paralabrax maculatorfasciatus: divergence of Gulf of California
and Pacific Coast populations. Evolution, 55, 1852-1862.
SWANK, S. (1988) Biochemical systematics of the genus Clinocottus (Cottidae). Bulletin of
the Southern California Academy of Science, 87, 57-66.
SWEARER, S. E., CASELLE, J. E., LEA, D. W. & WARNER, R. R. (1999) Larval
retention and recruitment in an island population of a coral-reef fish. Nature, 402,
799-802.
TAJIMA, F. (1989) Statistical method for testing the neutral mutation hypothesis by DNA
polymorphism. Genetics, 123, 585-595.
TAKAGI, M., CHOW, S., OKAMURA, T., SCHOLEY, V., NAKAZAWA, A.,
MARGULES, D., WEXLER, J. & TANIGUCHI, N. (2003) Mendelian inheritance
and variation of four microsatellite DNA markers in the yellowfin tuna Thunnus
albacares. Fisheries Science, 69, 1306-1308.
TERRY, A., BUCCIARELLI, G. & BERNARDI, G. (2000) Restricted gene flow and
incipient speciation in disjunct Pacific Ocean and Sea of Cortez populations of a reef
fish species, Girella nigricans. Evolution, 54, 652-659.
THOMAS, L. (1997) Retrospective power analysis. Conservation Biology, 11, 276-280.
THOMSON, C. (2001) Human Ecosystem Dimension. IN LEET, W., DEWEES, C.,
KLINGBEIL, R. & LARSON, E. (Eds.) California's living marine resources: A
status report., California Fish and Game.
TRANAH, G. & ALLEN, L. (1999) Morphological and genetic variation among six
populations of the spotted sand bass, Paralabrax maculatofasciatus, form southern
California to the upper Sea of Cortez. Bulletin of the Southern California Academy
of Sciences, 98, 103-118.
TURNER, T., WARES, J. & GOLD, J. (2002) Genetic effective size is three orders of
magnitude smaller than adult census size in an abundant, estuarine-dependent
marine fish (Sciaenops ocellatus). Genetics, 162, 1329-1339.
VAN HEYNINGEN, V. & YEYATI, L. (2004) Mechanisms of non-Mendelian inheritance
in genetic disease. Human Molecular Genetics, 13, 225-233.
175
VAN'T HOF, A., ZWANN, B., SACCHERI, I., DALY, D., BOT, A. & BRAKEFIELD, P.
(2005) Characterization of 28 microsatellite loci for the butterfly Bicyclus anynana.
Molecular Ecology Notes, 5, 169-172.
VOGEL, A., BURRELL, C., EDMANDS, S. & KIEFER, D. (in prep-a) Correlating
processes associated with ocean triads to the recruitment of the kelp bass,
Paralabrax clathratus.
VOGEL, A. & EDMANDS, S. (in prep-a) The influence of sample size and population
differentiation on statistical power and estimation of F
ST
in marine species.
VOGEL, A. & EDMANDS, S. (in prep-b) Population genetic structure in Paralabrax
clathratus and Clinocottus analis across potential barriers to gene flow.
VOGEL, A. & EDMANDS, S. (in prep-c) Temporal and spatial genetic variance in the kelp
bass, Paralabrax clathratus.
VOGEL, A., SELKOE, K., ANDERSON, D. & EDMANDS, S. (in prep-b) Development
and inheritance of molecular markers in the kelp bass, Paralabrax clathratus.
WAPLES, R. (1986) A multispecies approach to the analysis of gene flow in marine shore
fishes. Marine Biology. University of California, San Diego.
WAPLES, R. (1987) A multispecies approach to the analysis of gene flow in marine shore
fishes. Evolution, 41, 385-400.
WAPLES, R. (1989) A generalized approach for estimating effective population size from
temporal changes in allele frequency. Genetics, 121, 379-391.
WAPLES, R. (1998) Seperating the wheat from the chaff: patterns of genetic differentiation
in high gene flow species. Journal of Heredity, 98, 438-450.
WAPLES, R. & GAGGIOTTI, O. (2006) What is a population? An empirical evaluation of
some genetic methods for identifying the number of gene pools and their degree of
connectivity. Molecular Ecology, 15, 1419-1439.
WAPLES, R. S. & ROSENBLATT, R. H. (1987) Patterns of larval drift in southern
California marine shore fishes inferred from allozyme data. Fisheries Bulletin, 85, 1-
11.
WARD, R., WOODWARK, M. & SKIBINSKI, D. (1994) A comparison of genetic diversity
levels in marine, freshwater and anadromous fish. Journal of Fish Biology, 44, 213-
232.
WARD, R. D. (2000) Genetics in fisheries management. Hydrobiologia, 420, 191-201.
WEIR, B. & COCKERHAM, C. (1984) Estimating F-statistics for the analysis of population
structure. Evolution, 38, 1358-1370.
WRIGHT, S. (1931) Evolution in Mendelian populations. Genetics, 16, 97-159.
WRIGHT, S. (1951) The genetical structure of populations. Annals of Eugenics, 15, 323-
353.
YOSHIYAMA, R., GAYLORD, K., PHILIPPART, M., MOORE, T., JORDAN, J., COON,
C., SCHALK, L., VALPEY, C. & TOSQUES, I. (1992) Homing behavior and site
fidelity in intertidal sculpins (Pisces: Cottidae). Journal of Experimental Marine
Biology and Ecology, 160, 115-130.
YOSHIYAMA, R. & SASSAMAN, C. (1987) Geographical patterns of allozymic variation
in three species of intertidal sculpins. Environmental Biology of Fishes, 20, 203-218.
176
Appendix 1. Summary statistics for mitochondrial data. Sample size (n), allele number (a)
and haplotype diversity (h) are given. Letters in parentheses refer to sample locations in
Figure 22.
n a h
Locations
Abreojos (AB) 959 31 0.7454
Coronados (CO) 293 23 0.7743
San Diego (SD) 491 28 0.7743
San Clemente (CE) 287 25 0.7023
Santa Catalina (SC) 150 19 0.7627
Santa Barbara (SB) 137 21 0.7396
Recruits
Santa Catalina 1998 (SCR98) 91 16 0.7778
Santa Catalina 2004 (SCR04) 99 18 0.7724
Santa Catalina 2005 (SCR05) 99 13 0.7273
Carpenteria 2001 (CPR01) 86 15 0.7521
177
Appendix 1. Continued.
n a h
Location-year classes
Abreojos1975 1 1 1.0000
Abreojos1976 4 3 0.8333
Abreojos1979 3 3 1.0000
Abreojos1980 6 4 0.8000
Abreojos1981 7 5 0.8571
Abreojos1982 18 8 0.7516
Abreojos1983 6 3 0.6000
Abreojos1984 16 10 0.9167
Abreojos1985 11 6 0.7821
Abreojos1986 12 5 0.7424
Abreojos1987 19 7 0.7602
Abreojos1988 14 7 0.7630
Abreojos1989 220 21 0.7630
Abreojos1990 77 15 0.7485
Abreojos1991 25 7 0.6867
Abreojos1992 86 17 0.7382
Abreojos1993 20 7 0.7895
Abreojos1994 22 9 0.7056
Abreojos1995 40 12 0.7077
Abreojos1996 145 22 0.7233
Abreojos1997 134 20 0.7566
Abreojos1998 15 3 0.3619
Abreojos1999 25 10 0.8633
Abreojos2000 28 8 0.7302
Coronados1979 1 1 1.0000
Coronados1982 2 2 1.0000
Coronados1984 2 2 1.0000
Coronados1985 5 3 0.7000
Coronados1987 2 1 0.0000
Coronados1988 4 2 0.6667
Coronados1989 4 2 0.5000
Coronados1990 4 3 0.8333
Coronados1991 6 5 0.9333
Coronados1992 3 2 0.6667
Coronados1993 25 8 0.7267
Coronados1994 24 12 0.8949
Coronados1995 98 16 0.7871
Coronados1996 42 9 0.7259
Coronados1997 65 15 0.7803
Coronados1998 5 4 0.9000
Coronados1999 1 1 1.0000
178
Appendix 1. Continued.
n a h
SanDiego1985 1 1 1.0000
SanDiego1986 2 2 1.0000
SanDiego1987 1 1 1.0000
SanDiego1988 1 1 1.0000
SanDiego1989 8 3 0.4643
SanDiego1990 11 7 0.8182
SanDiego1991 13 2 0.4615
SanDiego1992 19 8 0.8421
SanDiego1993 84 18 0.7940
SanDiego1994 80 16 0.8228
SanDiego1995 137 19 0.7865
SanDiego1996 73 15 0.7664
SanDiego1997 55 15 0.7279
SanDiego1998 2 2 1.0000
SanClemente1969 1 1 1.0000
SanClemente1973 1 1 1.0000
SanClemente1974 1 1 1.0000
SanClemente1975 1 1 1.0000
SanClemente1977 2 2 1.0000
SanClemente1979 3 2 0.6667
SanClemente1980 3 3 1.0000
SanClemente1981 5 2 0.4000
SanClemente1982 12 8 0.8485
SanClemente1983 10 5 0.6667
SanClemente1984 8 3 0.5974
SanClemente1985 22 7 0.5974
SanClemente1986 10 5 0.7556
SanClemente1987 15 9 0.9048
SanClemente1988 14 5 0.5934
SanClemente1989 23 8 0.5810
SanClemente1990 46 11 0.7005
SanClemente1991 15 8 0.7905
SanClemente1992 10 3 0.5111
SanClemente1993 39 14 0.7773
SanClemente1994 8 3 0.6071
SanClemente1995 23 8 0.7510
SanClemente1996 9 4 0.6944
SanClemente1997 6 5 0.9333
179
Appendix 1. Continued.
n a h
SantaCatalina1988 1 1 1.0000
SantaCatalina1989 4 3 0.8333
SantaCatalina1990 2 2 1.0000
SantaCatalina1991 7 5 0.9048
SantaCatalina1992 6 4 1.0000
SantaCatalina1993 36 9 0.5441
SantaCatalina1994 8 3 0.6667
SantaCatalina1995 45 9 0.7379
SantaCatalina1996 22 10 0.9064
SantaCatalina1997 43 11 0.7857
SantaCatalina1998 8 3 0.7000
SantaCatalina1999 3 2 1.0000
SantaBarbara1991 5 4 0.9000
SantaBarbara1992 3 2 0.6667
SantaBarbara1993 18 6 0.7712
SantaBarbara1994 61 16 0.7525
SantaBarbara1995 11 3 0.4727
SantaBarbara1996 16 8 0.8500
SantaBarbara1997 23 6 0.6482
180
Appendix 2. Summary statistics for microsatellite data. Sample size (n), allele number (a)
and Nei’s unbiased heterozygosity (H) are given. Letters in parentheses refer to sample
locations in Figure 22.
n a H
Locations
Abreojos (AB) 929 179 0.7859
Coronados (CO) 306 142 0.7744
San Diego (SD) 530 166 0.7797
San Clemente (CE) 288 144 0.7780
Santa Catalina (SC) 187 141 0.7736
Santa Barbara (SB) 143 127 0.7710
Recruits
SantaCatalina1998 (SCR98)
SantaCatalina2004 (SCR04) 142 123 0.7789
SantaCatalina2005 (SCR05) 93 114 0.7562
Carpenteria2001 (CPR01) 93 114 0.7748
181
Appendix 2. Continued.
n a H
Location-year classes
Abreojos1975 1 12 0.7143
Abreojos1976 4 34 0.8061
Abreojos1979 4 36 0.7704
Abreojos1980 7 51 0.8148
Abreojos1981 7 48 0.7595
Abreojos1982 19 68 0.7736
Abreojos1983 6 39 0.7814
Abreojos1984 16 70 0.7834
Abreojos1985 11 68 0.7640
Abreojos1986 11 59 0.7727
Abreojos1987 17 75 0.8066
Abreojos1988 14 71 0.7834
Abreojos1989 220 142 0.7855
Abreojos1990 76 113 0.7812
Abreojos1991 26 84 0.7910
Abreojos1992 83 126 0.7866
Abreojos1993 17 80 0.7719
Abreojos1994 21 82 0.8025
Abreojos1995 36 100 0.7813
Abreojos1996 135 136 0.7815
Abreojos1997 123 131 0.7882
Abreojos1998 15 67 0.7875
Abreojos1999 28 93 0.7853
Abreojos2000 29 88 0.7917
Coronados1979 1 14 1.0000
Coronados1982 2 22 0.8333
Coronados1984 2 23 0.8571
Coronados1985 5 42 0.8095
Coronados1986 1 12 0.7143
Coronados1987 2 23 0.8810
Coronados1988 5 37 0.8095
Coronados1989 4 34 0.7806
Coronados1990 4 34 0.7602
Coronados1991 6 42 0.8074
Coronados1992 3 29 0.7905
Coronados1993 25 80 0.7901
Coronados1994 24 83 0.7920
Coronados1995 101 115 0.7673
Coronados1996 43 95 0.7679
Coronados1997 70 114 0.7664
Coronados1998 6 44 0.7727
Coronados1999 1 13 0.8571
182
Appendix 2. Continued.
n a H
SanDiego1985 1 11 0.5714
SanDiego1986 2 22 0.7619
SanDiego1987 1 12 0.7143
SanDiego1988 1 13 0.8571
SanDiego1989 9 52 0.7619
SanDiego1990 12 67 0.7888
SanDiego1991 13 65 0.7895
SanDiego1992 21 78 0.7675
SanDiego1993 89 117 0.7810
SanDiego1994 92 122 0.7677
SanDiego1995 147 135 0.7844
SanDiego1996 81 109 0.7799
SanDiego1997 55 110 0.7901
SanDiego1998 2 22 0.8333
SanClemente1969 1 11 0.5714
SanClemente1973 1 12 0.7413
SanClemente1974 1 13 0.8571
SanClemente1975 1 13 0.8571
SanClemente1977 2 21 0.8095
SanClemente1979 3 27 0.8000
SanClemente1980 4 37 0.8112
SanClemente1981 5 39 0.7524
SanClemente1982 12 66 0.7769
SanClemente1983 11 60 0.7848
SanClemente1984 8 48 0.7857
SanClemente1985 24 81 0.7880
SanClemente1986 10 59 0.8045
SanClemente1987 15 65 0.7718
SanClemente1988 13 64 0.7635
SanClemente1989 25 85 0.7826
SanClemente1990 44 104 0.7895
SanClemente1991 15 63 0.7929
SanClemente1992 11 59 0.7520
SanClemente1993 35 90 0.7604
SanClemente1994 8 49 0.7596
SanClemente1995 24 81 0.7748
SanClemente1996 9 49 0.7625
SanClemente1997 6 40 0.7721
183
Appendix 2. Continued.
n a H
SantaCatalina1986 1 9 0.2857
SantaCatalina1988 1 13 0.8571
SantaCatalina1989 4 35 0.8112
SantaCatalina1990 2 23 0.8571
SantaCatalina1991 7 42 0.7316
SantaCatalina1992 6 47 0.7727
SantaCatalina1993 36 89 0.7808
SantaCatalina1994 8 51 0.7917
SantaCatalina1995 45 99 0.7665
SantaCatalina1996 22 84 0.7779
SantaCatalina1997 43 94 0.7715
SantaCatalina1998 8 49 0.7676
SantaCatalina1999 3 26 0.7619
SantaBarbara1991 4 38 0.8037
SantaBarbara1992 3 30 0.8095
SantaBarbara1993 17 71 0.7502
SantaBarbara1994 63 110 0.7722
SantaBarbara1995 13 62 0.7797
SantaBarbara1996 18 70 0.7794
SantaBarbara1997 25 86 0.7732
Abstract (if available)
Abstract
Many longstanding questions in marine science center on understanding population variability. Research efforts that address these issues include analyses of recruitment and studies of genetic variance. Understanding recruitment variability is important in that it can change stock abundance and age structure, whereas genetic differences can be reflective of differential parental contribution, genetic drift and migration. This thesis is a report on variability in the recruitment and genetic structure in the kelp bass, Paralabrax clathratus, an important sport fishery in Southern California.
Linked assets
University of Southern California Dissertations and Theses
Conceptually similar
PDF
Genetic analysis of population structure in striped marlin, Tetrapturus audax, in the Pacific Ocean
PDF
Developing genetic tools to assist in the domestication of giant kelp
PDF
Simulated and field environmental effects on the transcriptome and metabolome of mussel Mytilus californianus
PDF
Multiple generations of hybridization between populations of the intertidal copepod Tigriopus californicus
PDF
Genetic architecture underlying variation in different traits in the Pacific oyster Crassostrea gigas
PDF
Genome-wide analysis of genetic load and larval mortality in a highly fecund marine invertebrate, the Pacific Oyster Crassostrea gigas
PDF
Transgenerational inheritance of thermal tolerance in two coral species in the Florida Keys
Asset Metadata
Creator
Vogel, Augustus
(author)
Core Title
Population genetics and recruitment of the kelp bass, Paralabrax clathratus
School
College of Letters, Arts and Sciences
Degree
Doctor of Philosophy
Degree Program
Biology
Publication Date
11/21/2006
Defense Date
10/11/2006
Publisher
University of Southern California
(original),
University of Southern California. Libraries
(digital)
Tag
genetic,OAI-PMH Harvest,otolith,Paralabrax,recruiment,temporal,triad
Language
English
Advisor
Edmands, Suzanne (
committee chair
), Hedgecock, Dennis (
committee member
), Kiefer, Dale A. (
committee member
), Michaels, Anthony (
committee member
), Thacker, Christine (
committee member
), Wilson, John P. (
committee member
)
Creator Email
avogel@usc.edu
Permanent Link (DOI)
https://doi.org/10.25549/usctheses-m184
Unique identifier
UC154139
Identifier
etd-Vogel-20061121 (filename),usctheses-m40 (legacy collection record id),usctheses-c127-37478 (legacy record id),usctheses-m184 (legacy record id)
Legacy Identifier
etd-Vogel-20061121.pdf
Dmrecord
37478
Document Type
Dissertation
Rights
Vogel, Augustus
Type
texts
Source
University of Southern California
(contributing entity),
University of Southern California Dissertations and Theses
(collection)
Repository Name
Libraries, University of Southern California
Repository Location
Los Angeles, California
Repository Email
cisadmin@lib.usc.edu
Tags
genetic
otolith
Paralabrax
recruiment
temporal
triad