offshore locations on the Newfoundland continental shelf. S. M. CARR, A. ...... lantic. International Commission for the Northwest Atlantic Fisher- ... Government.
Molecular Ecology (1995)4,7948
Mitochondria1 DNA sequence variation and genetic stock structure of Atlantic cod (Gadus morhua) from bay and offshore locations on the Newfoundland continental shelf S. M . CARR, A . J . SNELLEN, K. A. HOWSE' and J . S. WROBLEWSKI' Genetics, Evolution, and Molecular Systematics Lmboratory, Department of Biology, and, 'Fisheries Oceanography Group, Ocean Sciences Centre, Memorial University of Newfoundland, St. John's NF AIB 3x9, Canada
Abstract Bay cod, Atlantic cod (Gadus morkua) that over-winter in the deepwater bays of northeastern Newfoundland, have historically been regarded as distinct in migration and spawning behaviour from offshore (Grand Bank) cod stocks. To investigate their genetic relationships, we determined the DNA sequence of a 307-base-pair portion of the mitochondria1 cytochrome b gene for 236 adult cod taken from the waters off northeastern Newfoundland, including fish found over-wintering and spawning in Trinity Bay. Although 17genotypeswere found, a single common genotype occurs at a frequency of greater than 80% in all samples, and no alternative genotype occurs at a frequency of greater than 3%. Genotype proportions did not differ significantly among samples. Measures of genetic subdivision among sampling locations are nil. Cod over-wintering in Trinity Bay are not genetically distinct from offshore cod. In combination with tagging and physiological studies, these data suggest that there is sufficient movement of cod between bay and offshore locations to prevent the development or maintenance of independent inshore stocks. Adult cad that over-winter in Trinity Bay appear to represent an assemblage of temporarily nonmigratory fish that have become physiologically acclimated to cold-water inshore environments. The pattern of genetic variation in northern cod suggests a recent population structure characterized by extensive movement of contemporaxy individuals superimposed on an older structure characterized by a bottleneck in the population size of cod in the north-westem Atlantic. Keywords: 'bay cod', genetic stock structure, mtDNA cytodvome b sequences, northem cod complex, polymerase chain reaction Received 5 January 1994; revision accepted 26 May 1994
Introduction Atlantic cod (Gadus morkua) living off Labrador and north-eastem Newfoundland in North Atlantic Fisheries Organization (NAFO) Divisions 2J,3K, and 3L (Templeman 1962, 1956, 1979) collectively form a stock complex known as 'northern cod'. Northern cod have for more than 400 years supported the world's most extensive commercial fishery (Innis 19541, but have recently undergone a dramatic decline in biomass (Baird etal. 1991, 1992). A moratorium currently exists on commercial fishing. A major unknown in efforts to rebuild the fishery
Correspondence: Steve Cam. Tel. (709)737 4776. Fax (709) 737 4000. E-mail x a r r 63 kean.ucs.rnun.ca
(WorkingGroup on Cod Enhancement 1994) is the extent to which cod found in the major deep-water bays of northeastern Newfoundland represent separate stocks. Northern cod migrate each summer from offshore over-wintering grounds along the edge of the continental shelf to inshore areas along the coasts of Labrador and Newfoundland, where they feed upon capelin (Templeman 1966; Rose 1993). In the fall and early winter, the fish return offshore to the continental shelf, where they spawn during the following spring (Templeman 1979; Lear h Green 1984; Hutchings et af. 1993). The cycle is repeated annually, and the great majority of Northern cod conform to this pattern. However, some adult fish (so-called 'bay cod')over-winterin the larger, deepwater bays of northsastern Newfoundland (Thompson 1943; Wroblewski etal. 1994a). The migration and spawning
80 S. M. C A R R ef nl. behaviours of these fish with respect to offshore cod are poorly understood. It is recognized that bay cod have historically made a substantial contribution to the success of local inshore fisheries (Pinhorn 1984; Keats etal. 1986; Wroblewski et ul. 1994a). Clarification of the stock status of bay cod is a first step towards the improved management of northern cod. Jakobsen (1987) and God0 & Moksness (1987) have discussed similar research aimed at distinguishing coastal cod and offshore cod in Norwegian waters. A number of theories have been offered to explain the phenomenon of bay cod (Wroblewski et nl. 1994a). For example, bay cod may be an essentially random set of individual fish that participated in the summer inshore migration, and then in the fall remained behind while the majority of fish returned offshore. If the inshore-offshore migratory behaviour of individual fish varied from year to year, bay cod would be expected to have a high tumover rate and would not constitute distinct stocks. Alternatively, bay cod may be fish that remain more or less permanently inshore, with limited introgression of new individuals from offshore. Templeman (1979) believed that the location of adult cod during the winter/spring months was an indication of their spawning area; if so, bay cod would be expected to spawn inshore. Circumstantial evidence suggests that bay cod do spawn inshore. Fish in spawning condition have been reported from a number of inshore locations for many years: Graham (1922) cites evidence from as far back as 1889 of .the presence in Trinity Bay, Newfoundland, of cod that were extruding eggs and milt during spring and summer months. Based on such observations, Graham (1922) and later Thompson (1943) reasoned that cod that over-winter in Trinity Bay may also spawn there during the summer. Hutchings eta!. (1993) have reviewed temporal patterns of inshore and offshore spawning, and conclude that fish spawning in the bays are unlikely to be latespawning members of inshore migrating schools. If spawning were temporally as well as spatially distinct, bay cod might constitute discrete biological populations (Templeman 1966; Hutchings etal. 1993). However, there has been no direct test of this hypothesis. Movement patterns of cod over-wintering in Trinity Bay (Wroblewski et nl. 1994a,b) and their physiological acclimation to winter Ocean temperatures (Goddard et al. 1994) have recently been documented. Cod in nearspawning condition were observed in the Random Island region of Trinity Bay during the winter and spring months of 1990 and 1991 (Howse 1993). Cod in the act of spawning were observed over Heart's Ease Ledge in Trinity Bay during July 1993 (Wroblewski etal. 1994b). Genetic analysis of these fish in comparison with other fish taken from offshore spawning areas in the northern cod complex affords the first clear test of the degree of genetic stock separation among these groups of fish.
To be treated as separate populations, fish stocks must be genetically as well as ecologically distinct (HardenJones 1968; Utter 1991; Dizon et ul. 1992). Genetic differentiation in Atlantic cod has been examined by a variety of meaiv, including protein electrophoresis (Jamieson 1975; Cross & Payne 1978; Mork et al. 1985; Grant & Stdhl 1988a,b; Dahle & Jsrstad 19931, indirect analysis of mitochondrial DNA (mtDNA) sequences by restriction fragment length polymorphism (RFLP) (Smith el al. 1989; Dahle 1991; Amason et nf. 1992), and most recently by direct sequencing of mitochondria1 genes, amplified in uitro by the polymerase chain reaction (Carr & Marshall 1991a,b; Pepin & Carr 1993).The last provides an unusually sensitive method to explore differences among closely related individuals and populations within species. Here, we extend this analysis to a comparison of bay cod vs. offshore cod off north-east&n Newfoundland.
Materials and methods Cod were collected by commercial fishermen, personnel of the Canada Department of Fisheries and Oceans (DFO), and ourselves. Five localities in Northwest Atlantic Fisheries Organization (NAFO) Division 3L were sampled (Fig. 1). Sample sizes are listed in Table 1. The Northern Grand Banks sample (NGB) was collected in June 1988 as part of a DFO survey offshore cod. The Flatrock sample (FTK) was collected on a single day in June 1989 from a commercial fish processing plant on the east coast of the Avalon Peninsula north of St. John's; the
YL?
51
D h b n 3U
50
7
49
I -
Nonhern GnadBlmk
48
47
55
54
53
52
51
E
Fig. 1 Geographic distributionof cod samples in this study.
G E N E T I C STRUCTURE OF B A Y A N D OFFSHORE NORTHERN C O D
81
Table 1 Distribution of mtDNA genotypes of Gadus rnorhw among five sampling locations from north-eastern Newfoundland (NAFO Division 3L)
mtDNA genotype ~
Locality NGB FTK
ccv
RAN HEL Total
n
21 88 47 47 33 2 3
A
B
C
D
E
H
19 71
0 1
0 1
1 3
0 3
3
4 40 29 6
1 0 0 2
0
0 1 0
0
2
0
0
0 1
0
0
1
1
2
f
N
O
P
1 1
0 3
0 0
0 0
0 0 0 7
0
0 0
4
1 0 0 3
0
I 0 2
Codes for sample localities (see text for detail): NGB =Northern Grand Banks; RAN = Random Island,Trinity Bay; HEL = Heart's Ease Ledge,Trinity Bay.
sample consists mostly if not entirely offshore cod (Chen 1993). The Chance Cove sample (CCV)was collected in April and May 1990 from the inner part of Trinity Bay. The Random Island sample (RAN) was collected between January and April 1991 from the west side of Trinity Bay; the Heart's Ease Ledge sample (HEL)was collected near Random Island during July 1993. The CCV and RAN samples are representative of over-wintering bay cod. The HEL sample is representative of cod spawning in Trinity Bay. The RAN and HEL samples are from fish for which there are extensive tagging, sonic tracking. and physiological data (Wroblewski et al. 1994a,b; Goddard et al. 1994; see Discussion). Extraction of DNA from frozen heart tissue, symmetric and asymmetric amplification by the polymerase chain reaction, and DNA sequencing were performed substantially as described by Cam & Marshall (1991a). Some DNA sequences were obtained with the use of a fluorescent dyeterminator chemistry and an Applied Biosystems 373A Automated DNA sequencer (Can & Marshall 1991b). We used as amplification and sequencing primers a pair of oligonucleotides (L14841 and Hl.5749) that amplify a 359-bp portion of the mitochondrial cytochrome b sequences (Kocher e t d . 1989); the middle 307 bp of the amplification product will show naturally occurring genetic variation. Experiments on the HEL sample were done with primers L14724 and H15149 (Irwin eta!. 1991), which include an additional 94 bp of cytochrome b sequence. The primers were synthesized on a Milligen oligonucleotide synthesizer in the nudeic add analysis facility at Memorial University. All sequences are given as their coding strand equivalents. Sequences were analysed and prepared for publication with the help of the ESEE (version 2.00) program of Eric Cabot (Cabot & Bedcenbach 1989). Genetic heterogeneity within samples was estimated by the nucleon diversity (h) index (for nonselfing populations) and nucleotide diversity (x) index of Nei & Tajima (1981)(equations 8.4 and 10.4 of Nei 1987) as calculated by the REAP program
S
T
X
0
0 I
0 0
0 1
0
0
1 0 4
R
1
1 0 0 1
Y
0 0 1 1 0 0 0 0 1 2 1 0 0 0 1 3 1 2 2
~~~
Z 0 0
0 1 0 1
a 0 0
0 0 1
1
mK = Ratrock; CCV = Chance Cove, Trinity Bay;
(McEIroy et nl. 1991) from pairwise haplotype divergences calculated by the MEGA program (Kumar etal. 1993). The nucleon diversity index is approximately equivalent to the probability that two randomly chosen individuals will have different genotypes. The nucleotide diversity index measures the average pairwise nucleotide difference between individuals within samples, and corrects h for the size of the nucleon examined (Nei 1987). The heterogeneity of genotype distribution among samples was tested with the Monte-Carlo chi-square test of Roff & Bentzen (1989) as implemented in REAP, which is suitable for genetic data matrices in which many or most elements are very small (c5)or zero. A total of 5000 resamplings of the data matrix were used. The proportion of genetic diversity attributable to subdivision among samples was estimated by the coancestry coefficient (9) calculated with the HAPLOID program Weir 1990>,which measures the extent of genotypic differentiation among populations as the probability that two genotypes drawn at random from two different samples are identical by descent. For haploid genetic systems such as mtDNA, 9 is equivalent to the more familiar Ffl (Wright 1951)used for diploid nuclear genes. The standard error of 9 was estimated by jack-knifing over samples. We estimated the degree of gene flow among samples by the method of Slatkin & Maddison (1989), using a program supplied by M.Slatkin (Department of Integrative Biology, University of California, Berkeley, CA 94720, USA). This program estimates the product Np, where N, is the effective population size and m is the proportion of &grants exchanged among populations per generation. Confidence intervals of N*m were estimated by bootstrapping.
Results Within the middle 307 bp of the amplified segment, 16 variable positions were identified among 236 cod (Fig. 2). (No variable positions were observed in the additional 94-bp region examined in HEL sample.] All substitutions
82 S. M. C A R R et al.
A
B
c D
B E J
n
0
P R
s T x Y
z
a
P
G
S
L
L
G
L
C
L
I
!
Q
L
L
T
G
L
P
L
A
H
E
Y
T
S
D
2
6
at ttt gy tct ctt cta ggc ctt tgc tta att act ma ctt cta aca gga cta ttt cta gcc ata cac tat acc tca gac 80
................................................................................ ................................................................................ ................................................................................ ...................................................................... c . . . ...... ... ............... ......... ... ......... ... ............... ..... ............ ... ... ............... ..... ............... ... ... ......... ......... t.. ... ......... ... ............... ............ ... ............... ......... ,.. ..... ... ............ ... ... ... ......... ... ............... ..... ... ............ ............... ......... ... ... ... ............... ..... ... ............ ............... ......... ... ......... ... ... ......... ... ......... ... ............... ..... ... ............ ... ............... ... ......... ... ............... ... ..... ... ............ ............... ......... ... ... ......... ... ......... ... ............... ..... ... ............ ... ............... ... ......... ..... ... ............ ............... ... ... ..g ... ... ............... ... ... ......... ......... ... ... ......... ......... ... ... ............... ............... ... ............ ... ............... ... ..... ..... ... ............ ............... ... ..... ............ ... ............... ... ......... ... ......... ... ............ ..t ..I
.I.
D V I Y C Y L R l l H E n 53 gat a$, aac tac gqc t9a aat 161 cgg aat ata cat 4 ... ............... ............... ... ...... *.g ...... ......... ............ ... ... ...
I E T A P S S V V B I A atc gag aca gcc ttc tca toc gta gtc cac atc B
c
D B E
J II 0
P R s P I Y B
a
... ............... ............... ... ............... ......... ............ ... ... ... ... ..a ............ ............... ... ............... ......... ............ ... ... ... ... ............... ............... ... ............... ......... ............ ... ...
... ............... ......... ..t ... ... ........................ ............ ... ... ... ... ..a ............ ............... ... ............... ......... ............ ... .. ... ............ ... ... ............... ......... ... ... ............... ............... ............ ... ..a ... ... ............... ......... ... ............... ............... ............ ... ............... ......... ... ... ............... ............... ... ... ......... ............ ... ... ... ............... ... ............... ............... ... ......... ............ ... ... ............... ... ............... ............... ... ............... ......... ............ ... ... ... ... ............... ............... ... ............ ... ... ... ............... ......... ............... ............... ... ... ............ ... ... ... ... ............... ........................ *.. ............... ............... ... ... ......... ............ ... ... ............... ............... ... ............... ............... ... ........................ ............ ... ... ... *..
1
..*
m..
.I.
G A S P P P I C L Y H B I A R G L Y Y C S Y L P V E T 8 0 qgt gcc tct ttc ttt ttc a t t tqt ctt tat atq cac att qcc cqa ggt ctc tat tat qt tcc tat d t ttt $a qag am 242 I)
c
D E H J
I
0
P R s
P
x
Y 2 a
......... ............ ... .................. ......... ............... ............ ... ............ ......... .................. ......... ... ............... ... ............ ......... ......... ............ ... ............... .................. ............ ... ......... ............ ... .................. ......... ............... ............ ... ......... ............ ... ......... .................. ............... ............ ... ............ ,.q ... ......... ............ ... ......... ............... ............ ... ......... ............ ... .................. ......... ............... ... ............ ......... ......... ............ ... .................. ............... ... ............ ...... ..c ......... ......... ............ ... ......... ............... ............ ... ............ ... ......... .................. ......... ............... ............ ... .................. ..t ...... ............ ... ......... .............. ... ............ ..q ...... ............ ... .................. ......... ............... ............ ... ......... ............ ... .................. ......... ............... ............ ... ..c ......... ... ......... .................. ......... ............... ... ............ ..c ............ ............ ... .................. ......... ............ ... ......... ............... ............ ... .................. ......... ............... ......... o
U U I G V V L P L L V H H T S P V C Y V L tat gtc ctc A tga aac atc w9 Q t t g t C & k c t t tta gta ata ata a tx tct tfx 4ta
w ......... .. . ......... ... ............ ... .................. ......... c ... ..t ... ... ............ ... ............... ... ......... ......... .. D ... ..t ... ... ............ ... ............... ... ......... ......... .. B ......... ... ............ ... .................. ......... ......... .. 0 ......... ............ ... ............... ... ......... ......... .. J ... ..t ... ... ............ ... ............... ... ......... ......... .. n ......... ... ............ ... ............... ... ......... ......... .. 0 ... ..t ... ... ............ ... ............... ... ......... ......... .. P ......... ... ............ ... ............... ... ......... ......... .. R ......... ... ............ ... ............ ..t ... ......... ......... .. s ................................................................. T ................................................................. x ................................................................. Y ................................................................. e ................................................................. a ................................................................. 5
6..
F i g 2 Variation in DNA sequence of M u s morhun within a 307-base-pairregion of the mitochondria1cytochrome b gene. In each of the 17 genotypes, each nucleotide is identical to that in genotype A except where indicated. The top line gives the inferred amino acid sequence according to the IUB single-letter code. Numbers adjacent to the first and second lines indicate positionnumbers in the protein and nucleotide sequences.
GENETIC STRUCTURE O F B A Y A N D OFFSHORE N O R T H E R N C O D occur at the third positions in their respective codons and are silent, i.e. they do not result in amino acid substitutions. All except one are transitions. The variable positions define 17 distinct sequence genotypes among cod, which differ by between one and four nucleotide substitutions (Fig. 3). Genotypes A, B, C, D,E, H,J, N,P and X have been reported previously (Carr & Marshall 1991a; Pepin & Carr 1993); genotypes 0, R, S,T, Y, Z and ’a’ are new. The new sequences have been submitted to GenBank and assigned the accession numbers U09624UO9630; the additional 94 bp examined in the HEL sample are included in accession number U10356. Most genotypes differ from genotype A by single unique substitutions.Genotypes C,D, J and 0 form a dade defined by a shared nucleotide substitution at position 246; within this clade, genotypes D and J differ from genotype C by a shared substitution at position 86 (Fig. 2). Genotype A is the most common genotype in all samples (overall mean 8546, range 81-90%) (Tablel). The most common alternative genotype (D) occurs at an overall frequency of 3.0% and in all sample sites except RAN. Genotypes C, D, N, S and X occur in both inshore and offshore samples; genotype J occurs in both offshore samples only. The remaining genotypes o m r only in single samples in one or two individual fish. The Monte Carlo test indicates no significant differences of genotype distributions among the five samples (x2, = 61.02, P 10.59). There are no significant differences if the offshore and Trinity Bay samples are pooled and compared (NGB + FIX vs. CCV + RAN + HEL) (x2,, = 17.20, P 2 0.35). There are no significant differences among fish taken in Trinity Bay (x2, = 25.90, P 2 0.50). The nucleon diversity (h) and nudeotide diversity ( p ) indices are given in Table 2; data from Carr & Marshall (1991a) and Pepin & Carr (1993) are included for com-
I4
I41
Table 2 Haplotype (h) and nudeotide (rr) diversity indices within samples of Gadus m h u a from 15 sampling locations from the north-westemAtlantic and Norway
Locality
NAFO division
NGB
3L
CCV
3L 3L
m
RAN HEL STA
SPR
3L 3L 3K 3K
CEN BEL
3L 3L
BRI
3L 3L 30 3K 3Ps 3Pn
NEGB SEGB
GIs
rn IAM Norway
-
h
x
0.1812 0.3457 0.2379 0.2745 0.2275 0.2139 0.2333 0.4762 0.2483 0.3755 0.3810 0.3387 0.2961 0.2648 0.2637
0.001494 0.001591 0.001100 0.001100 0.001104
0.8460
o.Ooo721 0.000817 0.001796 0.002179 0.001376 0.001401 0.001988 0.001251 0.001556 0.001867 0.004980
Codes for sample localities as in Table 1. Additional localities kom Pepin & Carr (1993): STA = St. Anthony; SPR = Springdale; CEN = Cenheville; BEL = Bellevue Beach; BRI = Brigus South; NEGB = NE Grand Banks;SEGB = SE Grand Banks. Additional localities from Carr & Marshall (1991a): GIS = Grey Island Shelf; STP = St. Pierre Bank; IAM = he am Mom; Norway = Tmms0, Norway.
parison. Among the current samples, the FIX sample has the highest genetic diversity, and also the largest sample size. The mean (* SD)coancestry coefficient 8 over the five samples is -0.0016 0.0072,i.e. not significantly different from zero. This result indicates that the proportion of genotypic differentiation attributable to subdivision among samples is negligible. Maddison & Slatkin’s (1989) estimate of N.m is based on a count of the minimum number of migration events (S) necessary to account for the observed distribution of genotypes among multiple populations. For example, the occurrence of a genotype in n samples requires a minimum of S = n - 1 events. From the phylogenetic relationships shown in Fig. 3 and the sample sizes and distribution of 17 genotypes among the five samples given in Tablel, we c a l d a t e S=16, from which we estimate Nlm = 0.6 (95%confidence interval 0.2-1.1).
*
Discussion Fig. 3 Phylogenetic relationships of 17 cod genotypes and their distributionamong five samples The single minimum-lengthnetwork is shown (length= 16). Each branch represents a single nucleotide substitution. Numbers in brackets indicate samples in which each genotype is found. Codes for samples: 1 = NGB, 2 = FW,3 = CCV,4 RAN.5 = HEL
83
-.
Genetic and ecological differentiation of buy cod and oflshore cod Adult cod in NAFO Division 3L show extensive genetic polymorphism (17 genotypes) but low heterogeneity (aggregate h = 0.30; aggregate R = 0.001341). Comparison of
84 S. M.C A R R et nl.
cod over-wintering in Trinity Bay with offshore cod indicates that genotype proportions are not significantly differentiated among samples. Cod found spawning in Trinity Bay are not differentiated from other fiih. There is no indication that such genetic variation as exists is subdivided among geographic samples (0 = 0.0).Estimates of gene flow among sampling locations are very low (Nem= 0.61,but this value may be a spurious function of other aspects of the population biology of these cod (see below). Intermixing between bay and offshore cod a p pears to be sufficient to produce a genetically homogeneous population of northern cod in this portion of the Newfoundland continental shelf. The absence of genetic differentiation is consistent with other data that indicate substantial mixing of bay and offshore cod. Tag-recapture data indicate that, although most cod tend to migrate annually between particular ofkhore spawning grounds and inshore feeding grounds, many individuals stray from this movement pattern (Templeman 1979; Lear 1984). Recent tagging studies of adult cod in Trinity Bay show that some cod remain there for more than one year, others undergo long-distance movements from bay to bay, and still 0thers apparently return offshore (Wroblewski et al. 1994a; Goddard ef at. 1994). Thus, the composition of any bay cod group is probably not sufficiently constant from year to year to allow it to develop as a reproductively independent separate stock: a very small amount of gene flow would be suffiaent to prevent such local genetic differentiation (Wright 1951; Slatkin 1985). All adult northern cod have the physiological capacity to over-winter in the bays (Goddard et nl. 1994). The ability of cod to over-winter in subzero inshore water is dependant on physiological acclimation.Cod generally prefer water temperatures between 0 and 5°C (Rose & Leggett 1988).Below the cold intermediate layer (300m), offshorewater temperatures remain above 0 "Call winter (Petrie et nl. 1988); inshore temperatures typically fall below 0°C in February (Wroblewski e l a l . 1993). Cod acclimate to subzero water by the production of antifreeze glycoproteins that lower the plasma freezing point to as low as -1.5 "C (Heweta!. 1981). The rate of induction is sufficiently rapid that cod remaining behind in the bays can acclimate to the decline in temperature over the winter months (Fletcher ef nl. 1987; Goddard et al. 1994). Goddard et nl. (1994) showed that blood glycoprotein levels could be used to measure how long fish had been exposed to subzero temperatures. Thermal hysteresis profiles of cod taken from the Random Island region of Trinity Bay during the spring of 1991 (the RAN sample) indicate that these fish had been exposed to subzero water for several months, and had likely over-wintered there. Profiles of the July 1993 inshore spawning cod (the EEL sample) showed no antifreeze glycoprotein activity
m
m
Fig. 4 Phylogenetic relationships of 22 cod genotypes from this study,Carr&Marshall(199la),and Pepin tkCarr(1993),and their distribution in the North Atlantic. One of two minimum-length trees is shown (length = 23): genotypes M and Y have identical nucleotidesubstitutions at position 221 in Fig. 2, which are shown here as having occurred in parallel. Each branch represents a single nucleotide substitution, except that between genotypes C and M which represents threesubstitutions.Numbers in brackets indicate localitiesin which each genotype is found. Codesfor localities (and groupings) from Table 2 1, NGB/FI'K (Division3L adult); 2, CCV/RAN/HEL cnnity Bay adult);3, STA/SPR (Divlsion3Kjuvenile);4, CEN/BEL/BRI/NEGB (Division3L juvenile); 5, SEGB; 6, GIs; 7, Srp; 8, IAM; 9, NOWJY.
(Wroblewski et al. 1994b), as expected for fish sampled in the summer. These cod may have over-wintered in Trinity Bay, or alternatively may have migrated into the bay from the offshore region (Lear etal. 1986) b u t see Hutchings et d . 1993). Bay cod in Trinity Bay thus seem to represent a transient assemblage of genetically indistinct fish that have moved into the bay and undergone a physiological acclimation that permits them to remain inshore while other fish return offshore. The factors that determine which fish migrate and which remain in the bay have yet to be studied.
Population genetics of Atlantic cod Recent studies of mtDNA have revealed sharply contrasting levels of genetic variability and population differentiation within and among sampling locations of cod across the North Atlantic. Smith etal. (1989) fgund no variation at all among a limited number of RFLP markers in a single sample of 16 cod from the Grand Banks. In contrast, Arnason et a!. (1992) found extensive variation in eight samples from Iceland (h =0.92 as calculated by Pepin & Carr 1993). Substantial RFLP variation has also been reported among Arcto-Norwegian cod (Dahle 1991). Previous studies of mitochondria1 cytochrome b sequences show a qualitatively similar pattern. Although multiple genotypes are observed among both adult and juvenile cod elsewhere in the northern cod complex and adjacent waters off Newfoundland and Labrador, geno-
G E N E T I C S T R U C T U R E O F BAY A N D O F F S H O R E N O R T H E R N C O D type A predominates in all samples, no alternative genotype occurs at an overall frequency of more than a few percent, and most genotypes are observed only once (Can & Marshall 1991a,b; Pepin & Can 1993). Among seven samples of juvenile cod from nursery areas in NAFO Divisions 3K, 3L, and 30, genetic diversity is low (aggregate h = 0.33, aggregate x = 0.001491) and the coancestry coefficient 8 is 4.0122; there is no indication of differentiation among samples (Pepin & Carr 1993). There are no significant differences between the samples reported here and those in previous studies; estimated pairwise nucleotide divergences between samples are very small and in many cases less than zero, indicating that within-sample divergence is greater than amongsample divergence. Mitochondria1 DNA differentiation among sampling locations in the Northern cod complex generally appears to be nil. In contrast, nucleotide variation in the eastern Atlantic appears to be substantial. A single sample from Norway showed greater genetic diversity (h = 0.85, x = 0.004980) than all western Atlantic samples combined, with no single genotype predominating and several coexisting in the same sample at moderate frequencies (Cam h Manhall 1991a). Icelandic cod also show extensive withinsample variation in the qtochrome b region examined here (Einar Arnason and coworkers, personal communication). Gene flow among populations (estimated by N p ) tends to reduce the degree of genetic subdivision among populations (measured here as 8). Low levels of gene flow ought therefore to be associated with high levels of genetic subdivision. For mtDNA, a value of N.m = 1.O is equivalent to the exchange of one individual per pair of samples per generation, which is roughly the lower limit of gene flow below which populations would begin to diverge by genetic drift alone (Slatkin 1985). The value of N . p = 0.6 obtained here is thus unexpected in conjunc tion with the essentially zero value of 8. Such a contrast has been observed previously in cod (Pepin & Carr 1993). The answer may lie in contemporary patterns of gene flow superimposed on a past history of population size fluctuation among cod in the western Atlantic, aspects of what Avise (1994) has termed ’shallow’ and ’deep‘ population structure.
‘ShalIow’ population structure in Northern cod The occurrence of ‘private alleles‘ (alIeles that occur only in a single population) is an important indication of reduced gene flow among populations (Slatkin 1985). The question of whether a particular allele is ’private’ becomes statistically problematic when within-sample genotype diversity is low. Slatkin & Maddison‘s (1989) algorithm does not take into consideration the frequency distribution of rare ‘private’ alleles: genotypes observed only once must by definition be ’private’, yet are not con-
85
vincing evidence of limited gene flow. Intuitively, the observation that six of nine rare genotypes occurring in at least two fish are shared between bay and offshore samples (Table 1) suggests that newly arisen genotypes can be rapidly exchanged among populations. Also, combined analysis of all 22 cod cytochrome b genotypes published to date (Cam & Marshall 1991a; Pepin h Carr 1993)shows that eleven genotypes (B-E,G, H,j, N,0, P,X ) which in separate studies appear rare and localized are in fact more widely distributed geographically (Fig. 4). All but one of the six rarer genotypes identified in the single Norwegian sample (genotypes 8-E, G) have now been found in the western Atlantic. Genotypes H, 0, and P,which occur in the present study only in one sample each, also occur in adult cod from NAFO Division 3K (H,0)and in juvenile cod from NAFO Division 3L (P).Genotypes C,E,and J occur in other NAFO divisions, and genotypes N and X occur in juvenile cod in Division 3L. As sampling becomes more comprehensive, the overall picture in Northern cod is one of extensive genotypic mixing and little differentiation. Low mtDNA differentiation over extensive geographic areas is characteristic of marine fish with planktonic larvae (Ovenden 1990). In contrast, comparisons of the distribution of distinct ’private‘ mtDNA alleles between northern cod and fish from the Gulf of Maine WAF0 Division 5Y) indicate significant genetic differentiation on either side of the Laurentian Channel (S. Carr, work in progress).
’Deep’ population structure in Atlantic cod The combination of extensive genotype polymorphism and low genetic diversity within the cod population of the western North Atlantic is consistent with the theory that this population has experienced a severe reduction in population numbers at some time in the past. Such ‘bottlenecks’ have a particularly severe impact on maternally inherited genes such as mtDNA (Wilson etal. 1985); in extreme cases, only a single maternal lineage or genotype may survive. When the population begins to increase in numbers after passing through such a bottleneck, all new genotypic variants will be derived from the common surviving lineage, and the result is a ’star phylbgeny’ with many rare genotypes differing from a common genotype by one or a few unique nucleotide substitutions (Slatkin & Hudson 1991). In such a phylogeny, relationships among individuals are poorly resolved because most branches are recent and therefore short, the minimum count of migration events Scan never be large, and there fore the actual extent of migration may be seriously underestimated (see discussion in Pepin & Carr 1993). Slatkin &K Maddison (1989) emphasize the need for a wellresolved phylogeny for meaningful quantification of gene flow by their approach. A ‘star phylogeny’ is apparent for cod from the western Atlantic (Figs3 and 4; cf.
86 S.M.CARR e t a i . figure 2 in Can & Marshall 1991b and figure 7 in Pepin & Carr 1993). Low contemporary estimates of gene flow in Northern cod may therefore be artefacts of this historic 'deep' structure. Actual gene flow may be more extensive, as suggested above. The hypothesized bottleneck might be a relatively ancient event associated with the Transatlantic dispersal of cod from the Old to the New World several millions of years ago (Grant & StAhl 1988a,b), a more recent event such as the last retreat of glacial ice from the Grand Banks approximately 10 OOO years ago (Cross & Payne 19781, or a more immediate or ongoing phenomenon associated with past and present fisheries practices in the last 100 years (Lear et aJ. 1986; Keats et al. 1986).One approach to distinguishing among these alternatives is to calculate the time of coalescence for the observed nucleotide diversity, that is, the estimated time since the mtDNA molecules in the sample last shared a common ancestor (Hart1 & Clark 1990). Nei (1985) estimated the nucleotide divergence rate for the entire mtDNA molecule in humans as 7.1 x lo9 per nucleotide site per million years per pair of lineages. The portion of the cytochrome b molecule examined here evolves slightly more quickly than the molecule as a whole: Carr & Hughes (1993) calculated a 75% sequence divergence vs. a 6.0% sequence divergence calculated from RFLP data for the molecule as a whole for two deer (Odocoilews) genotypes. [Analogous data for cod are not available; the 'molecular clock' in fish may be somewhat slower (Kocher et al. 1989)J.Given a nucleotide diversity of c. 0.0013 for Northern cod and a corrected cytochrome b divergence rate,
(0.0013)/ (7.1 x lW((1.2.5) = 1.5 x lo-' indicates a most recent common ancestor at c. 150000 years. The variance on this estimate is however, considerable; neither the Ice Age nor the trans-Atlantic migation hypotheses can be ruled out. It should be remembered that coalescence does not require or imply a severe bottleneck in population numbers (cf. Cann et nI. 1987). For haploid, maternally inherited genetic systems such as mtDNA, the estimated coalescence time is equivalent to the effective population number ( N ) , corrected by the number of years per generation (Hard& Clarke 1990). From the above estimate, a generation time of c. 5 years in cod gives N* = 30 OOO for Northern cod collectively. This is considerably less than the contemporary census size of c. 1oY (Baird et al. 19921, which may indicate that cod populations are still feeling the genetic effects of a relatively ancient bottleneck (Nei et al. 1975). The bottleneck theory is also consistent with the absence of marked allotyme differentiation among cod populations in the western North Atlantic (Cross & Payne 19781, as previously discussed (Can & Marshall 1991a).
Management implications of population genetic studies It has for many years been assumed that subcomponents of the Newfoundland cod fishery associated with various geographic subdivisions in the area constitute distinct 'stocks' for management purposes (reviewed by deYoung & Rose 1993). In the present case, bay cod are frequently alleged to constitute distinct 'bay stocks', independent from offshore stocks, based on their inshore over-wintering behaviour and summer spawning. Much effort has been devoted to looking for techniques, including rneristia (Templeman 198l), morphometrics (Pepin & Carr 19931, tagging (Lear 1984, 1986), parasites (Kahn et al. 1980; Brattey et al. 19901, and recently genetics, that would confirm the assumption of stock separation and so provide a tool for monitoring these and other 'stocks' within the northern cod complex. Where differences are not found, one might be tempted to conclude that the various techniques 'lack sensitivity' or have not been 'successful' because the results do not conform to expectation. An alternative approach is to measure such variation as exists, partition the observed variance among samples drawn from different pulative 'stocks', and based on these data to consider the alternate hypothesis that distinct cod stocks do not exist within the complex (Pepin & Carr 1993). This alternate hypothesis has important management implications. For example, proposals to rebuild 'bay stocks' by releasing large numbers of farmed (cage raised) cod into deepwater bays on the north-east coast of Newfoundland (Working Group on Cod Enhancement 1994)rely on the assumption that such fish would tend to remain inshore after release: behavioural, physiological, and genetic data suggest otherwise. In recovery strategies for restoration of northern cod, it is important to consider the genetic structure of the complex, and the cause(s1 of low mitochondria1 genetic variability in the complex. Comparisons should be made both with other cod populations, as well as with other genetic markers (Brooker eta!. 1994).
Acknowledgements We thank Einar Amason, T. Burke, P. Pepin, and two anonymous reviewen for discussion and critical comment. Einar and members of his lab at the University of Iceland discussed their results with us in advance of publication. R. Hudson and D. McElroy provided advice on statistical procedures. We thank R. Payne, K. Smedbol, E. Smith, and personnel of the Canada Department of Fisheries and Oceans (DFO),St. John's Branch, for co-operation in sample collection. We thank H. D. Marshall, E. Stewart, A. Greenslade, D. Crutcher, D. Evansand W. Bailey for expert technical assistance. This research was supportedby a contract from the Northem Cod Science Program (DFO)and grants from the Canadian Centre for Fsheries Innovation, the Natural Sciences and Engineering Research Council (NSERC),the DFO / NSERC Science Subvention Program, and the Ocean Production Enhance-
GE NETIC STRUCTURE O F B A Y A N D OFFSHORE N O R T H E R N COD ment Network (Networks of Centres of Excellence) to SMC, and funds from National Sea M u c k and Fishery Products International under the NSERC Research Partnership Program to JSW.
References Arnason E, Palsson S, Arason A (1992) Gene flow and lack of population differentiation in Atlantic cod, Gadus morhun L., from Iceland, and comparison of cod from Norway and Newfoundland. Journal ofFish Biology, 40,751-770. Avise, JC (1994) Molecular M a h , Natural History and E w l u t i a . Chapman &Hall, New York. Baird JW, BishopCA, BrodieWB, Murphy EF (1992)An assessment of the cod stock in NAFO divisions 2J3KL. Northwest Atlantic Fisheries Organization Scientific Committee Report 92/18, Ser. No. N2063. Baird JW, Bishop CA, Murphy EF (1991) Sudden changes in the perception of stock size and reference catch levels of cod in northeastern Newfoundland shelves. NAFO scient$c Cou?ISd Studies, 16,111-119. Brattey J, Bishop CA, Myers RA (1990) Geographic distribution and abundance of Pseudotmnoaa dm'piens (Nematoda: Ascaroidea) in the musculature of Atlantic cad, Gndus w h m , from Newfoundland and Labrador. In: Population biology of the sealworm ( P s e u d o t m m dm'piens)in relation to its intermediate and seal hosts (ed. Bowen WD). Camdim1 Bulktin of Fisheries and Aquatic Sciences, 222,674.. Brooker AL, Cook D, Benken P, WrightJM, Doyle RW (1994) The genomicorganization of microsatellitesdiffers between mammals and cold water teleost fishes. CanadinnJournnlofFishm'es and Aquatic Sciences, in press. Cabot EL, Beckenbach AT (1989)Simultaneousediting of multiple nudeic acid sequences with ESEE. Computer Applications in the BioSciences,5,233-234. CannRL, StonekingM,Wilson AC (1987)Mitochondrial DNA and human evolution. Nature, 3?5,31-36. Carr SM,Hughes GA (1993) The direction of hybridization between species of north American deer ~ O d m i l n c sas ) inferred from mitochondria1 cytochrome b sequenceS. Jouml of M m n d ~ g y74,331-342. , Carr SM, MarshallHD (1991a) Detection of intraspecificDNA sequence variation in the mitochondria1 cytochmme b gene of Atlanticcod (Gadusd u n ) by the polymerase chain reaction. Canadian Iournal of Fishmks and Aquatic Scicncrs. 48,48-52. Carr SM,Marshall HD (1991b)A direct approach to the measurement of genetic variation in fish populations: Applications of the polymerasechain reaction to studies of Atlanticcod (Gndus morhua). Journal ofFish Biology, 39 (Suppl. A), 101-107. Chen,X (1993)Sptinland trmporalpttemof landingwriation in the Newfoundlad northern cod inshore fishery. MSe thesis, Memorial University of Newfoundland. Cross TF, Payne RH (1978) Geographic variation in Atlantic cod, Gadus morhua, off eastern North America: a biochemical S p tematics approach. Journal of the Fishnics Research Board of Canada,35,117-123. Dahle G (1991) Cod, Gadus m h u a L., populations identified by mitochondrial DNA. Journal of Fish Biology,38,295-303. Dahle G, Jmtad KE (1993)Haemoglobin variation in cod: a reliable marker for arcticcod (cadusmorhw).FisheriesResearch, 16, 301-311. W o n AE, Lodcye C, Penin WF, Demaster DP, S i n J (1992)Rethinking the stock concept:a phylogeopphic approach. Conservation Biology,6,2446. deYoung B, Rose GA (1993) On recruitment and distribution of
87
Atlantic cod (Gadusmorhua) off Newfoundland. C a d i a n l o u r nal of Fishmk and Aquatic s r i e n c ~50,2729-2741. , Kao MH (1987) Low temperature regulation Fletcher GL, King MJ, of antifreeze glycoprotein levels in Atlantic cod (Gadus morhw). Canadinn ~OurnalofZoology,65,227-233. Goddard SV, Wroblewski JS, Taggart CT etal. (1994) Overwintering of adult Northern cod (Gadusmorhua) in cold inshore waters as evidenced by plasma antifreeze glycoprotein levels. Canadian IournaI ofFisheriesand Aquatic Sciences, in press. God0 OR, Moksness E (1987) Growth and maturation of Norwegian coastal cod and northeast Arctic cod under different conditions. Fisheries Ratnrch, 5,235-242. Graham M (1922)Invcstigations mrnprising o b s m t i o n s on the natural history ofthe Nnvfoundland shore cod. Report to the Director of Fshery Investigations,Fisheries Laboratory, Lowestoft, UK. Grant WS, Stdhl G (1988a) Evolution of Atlantic and Pacific cod: loss of genetic variation and gene expression in Pacific Cod. Evolution, 42,138-146. Grant WS, Stdhl G (1988b) Description of electrophoretic loci in Atlantic cod, Gudus morhua, and comparison with Pacific cod, Gadus mncnxephahs.Herditas, 108,27-36. Harden-JonesFR (1968) Fish Migration. Edwin Arnold, London. H a d DL, Clark AG (1990) Principles of Population Genetics, 2nd edn. Sinauer,Sunderland, MA. Hew CL, Slaughter 0,FletcherGL, JoshiSB (1981)Antifreeze production, freezer resistance, and overwintering of juvenile northern Atlanticcod (Gadusmorhua). CanndinnJournalof Zoology, 59,2186-2192 Howse KA (1993)A modified migration mudelfor Atlantic md, Gadus murhua, offthe northcast m t of Newfoundland. MSc thesis, Memorial University of Newfoundland. Hutchings JA, Myers RA, Lilly GR (1993)Geographic variation in the spawningof Atlanticcod, Wus morhua. Cnndian Journnlof Fishm'es and Aquatic Sciences, 50,2457-2467. Innis HA (1954) The Cod Fisheries, revised edn. University of Toronto Press, Toronto. Irwin DM, Kocher TD, Wilson AC (1991) Evolution of the cytochmme b gene of mammals. Journal of Molecular Evolution, 32,128-144. Jakobsen T (1987) Coastal cod in northern Norway. Fisheries Research, 5,223-234. Jamieson A (1975) Enzyme types of Atlantic cod stocks on the North American Banks. In: lsozymes, Vol. IV. Genetics and Em l u t i a (ed.Markert CL), pp. 491-515. Academic Press, New
Yo* Kahn RA,Muphy J,Taylor D (1980)Prevalenceof a trypanosome in Atlantic cod (Gadus morhua) especially in relation to stocks in the Newfoundland area. G M d i a n Journal of Fisheries and Aquatic SEienccs, 37,1467-1475. Keats D,Steele DH, Green JM(1986)A review ofthe recent status of the Northern Cod s t d (NAFO Divisions 2J, 3K, u@ 3L) and the declining inshorefishcry. A Report to the Newfoundland Inshore Fsheries Assodation, on Scientific Problems in the Northern Cod Controversy. [Availablefrom the Institute of Newfoundland Studies, Memorial University of Newfoundland, St. John's.] Kocher TD, Thomas WK, Meyer A et al. (1989) Dynamics of mitochondrial DNA evolution in animals: Amplification and sequencingwith conserved primers. Proceedings ofthe National Aardmry of Scienczs of the USA, 86,61964200. Kumar S,Tamura M,Nei M (1993)MEGA: Molecular Evolutionmy Genetics Analysis,Vexsion 1.01. Pennsylvania State University, CollegePark. b a r WH (1984) Discrimination of the stock complex of Atlantic cod (Gadus morhua) off southern Labrador and eastern New-
88 S.M.C A R R et nl. foundland, as inferred h m tagging studies. lournal of Northwest AtlanticFishnis Sciences, 5,143-159. Lear WH (2%) A further discussion of the stock complex of Atlantic cod (Gadus morhua) in NAFO divisions 21,3K, and 3L Northwest Atlantic Fisheries Organization Scientific Committee Re-
port,~oc.%/iia. Lear WH, Baird JW, Rice JC et al. (19%) An examination of fnctors amting atch in the inshore fishery of lnbrador and eastern Newfoundland. Canadian Technical Reports in Fisheries and Aquatic Sciences No. 1469. LearWH,Green JM(1984)Migrationof the’northern’ Atlanticcod and the mechanisms involved. In: Mechanisms of Migration in Fishes (eds McCleave JD eta/.), pp. 309-315. Plenum, New
York McElroy D, Moran P, Bermingham E, Kornfield I (1991)REAP: The Restriction Enzyme Analysis Package, Version 4.0. University of Maine, Orono. Mork J, Ryman N, Stdhl G, Utter FM,Sudnes, G (1985) Genetic variation in Atlanticcod (Gadus morhua) throughout its range. Canadian ~ournnlofFishcricsandAquaticSn’mas,42,1~1587. Nei M (1985) Human evolution at the molecular level. In: Population Genetics and hk&cular Ewlution (edsOhta T, Aoki K), pp. 41-64. Japanese Scientific Societies Press,Tokyo. Nei M (1997)Molecular Edutionury Genetics. Columbia University Ress, New York Nei M,Maruyama T, C h a h b o r t y R (1975)The bottleneck effect and genetic variability in populations. Ewlution, 29,l-lO. Nei M,Tajiia F (1981)DNA polymorphism detectable by restriction e n d o n u d e a s s Genetics,97,145163. Ovenden, JR (1990)Mitochondria1 DNA and marine stock assessment: a review. Australian Journal of Marine and Freshwater Research. 41,835-853. Pepin P, Carr SM (1993) Morphological, meristic, and genetic analysis of stock structure in juvenile Atlantic cod (Gadus morhua) h m the Newfoundland Shelf. Canadian lournal ofFishcries and Aquutic Scimctb, 50,1924-1933. Petrie 8, Akenhead S, Lzier J, Loder J (1988)The cold intermediate layer on the Labrador and northeastern Newfoundland shelves, 1978-1986 NAFO Science Council Studies, 12,5749 Pinhorn AT (1984) Inshore exploitation of Atlantic cod, Gadus m h u a , in Labrador and eastern Newfoundland waters. lourMI of Northwest AtlanticFishq Science. 5,79-84. Ro~DA,BentzanP(1989)Iheslatisficalanalysisofmitochondria1 DNA polymorphisms: x* and the problem of small samples. Molecular Blorogy and Ewlution, 6,539-545. Rose GA (1993) Cod spawning on a migration highway in the north-west Atlantic. Nature, 366,458-461. Rose GA, Leggett WC (1988)Atmosphere-ocean coupling and Atlantic cod migrations:effects of wind-forced variations in sea temperatures and currents on nearshore distributions and catch rates of Gadw morhua. Gmndian Journal of Fishmk and Aquaticsfimccp, 45,1234-1243. SlatMn M (1985) Gene flow in ~ h ~populations. ~ a l Annual Review 0fEcology and Systematis, 16,393-430. Slatkin M, Hudson RR (1991) Pairwise comparisons of mitochondrial DNA sequences in stable and exponentially growing populations. Genetics, 123.555562. Slatkin M, Maddison W P (1989) A cladistic measure of gene flow inferred from the phylogenies of alleles. Genetics, 123,60-13. smith PJ, Birley AJ, JamiesonA, Bishop CA (1989)Mitochondrial DNA in the Atbntic cod,Gudus morhua: lack of genetic divergence between eastern and western populations. lournal of Fish
Biology, 34,369-373. Templeman W (1962) Division of cod stocks in the Northwest Atlantic. International Commissionfor the Northwest Atlantic Fisheries Research Redbook, 1962 (III),79-129. Templeman W (1966)Mnrine Resources ofNewfoundland. Fisheries Research Board of Canada Bulletin No. 154. Templeman W (1979) Migrations and intermingling of stocks of Atlantic cod, Gadus morhua, of the Newfoundland and adjacent areas from tagging in 1962-66.IntrrnationaIh i s s i o n for the Northwest Atlantic Fisheries Rarnrch Bulletin, 14,550. Templeman W (1981) Vertebral numbers in Atlantic cod, Gadus morhuu, of the Newfoundland and adjacent areas, 1947-1971, and their use in delineating cod stocks. 1ouml of Northwest At1anticFishq Science, 2,2145. Thompson H (1943)A biological and economic study of cod (Gadus callarias L.) in the Newfoundland area. Research Bulletin 14 (Fisheries), Department of Natural Resources, Newfoundland Government. St. John‘s, Newfoundland. Utter F (1991)Biochemical genetics and fishery management: an historical perspective. Journal of Fish Biology, 39 (Suppl. A), 1-20.
Weir BS (1990) Intraspecific differentiation. In: MokfcularSystematics (eds Hillis DM, Moritz C), pp. 373410.Sinauer, Sunderland, MA. Wilson AC, Cann RL, Cam SM at al. (1985) Mitochondrial DNA and two perspectives on evolutionary genetics. Biolc@al Journal of Linnaean Society, 26,37SK!O. Working Group on Cod Enhancement (1994) Cod enhancement o p t i o n a l plans. Report of the urorking group on cod enhancement Draft #2. MS. [Available from the Canadian Centre for Fisheries Innovation, Memorial University of Newfoundland, St. John’s.] Wright S (1951) The genetical structure of populations. Annals of Eugenics. 15. 323-354. Wroblewski JS, Collier A, Bailey W (1993)Ocean temperaturein the Random Island Region of Trinity Bay, 1991-1992. The NSERC/ Fishery products International/Nationa1 Sea Roducts Chair in FEheries Oceanography, Data Report No. 1.Memorial University, St. John’s. Wroblewski IS, Bailey WL, Howse KA (1994a) Observations of adult northern cod (Gadus morhua)overwinteringin nearshore waters of Trinity Bay, Newfoundland. Canadian lournal of Fisheries and Aquatic Sciences, 51,142-150. Wroblewski IS,Bailey WL., Smedbol RK (1994b)B e h i o r offanned cod released in Trinity Bay, Newfoundland 1992-1993. The NSERC/Fishery Products Intemational/National Sea Products Chair in Fisheries Oceanogaphy, Data Report No. 2. M e morial University, St. John’s.
This paper is part of an ongoing collaborative effort by-Canadian geneticists, physiologists, and oceanographers to understand the biology of Atlantic cod with a view towards recovery and better managementoftheresource.SteveCarrisinterestedindifferences among the populationgenetic structuresof a variety of marineand terrestrial vertebrate speciesas revealed by molecular markers, an interest first developed in the laboratory of the late Allan Wilson. JoeWroblewski holds the NSERC /Fishery Products Intema tional/ National Sea Products Research Chair in Fisheries Oceanography at Memorial University.
