atlantic cod stock structure feature in the gulf of maine€¦ · national marine fisheries service...

fishe

ries

rese

arch

featu

re

10 Fisheries | www.fisheries.org | vol 29 no 1

The Gulf of Maine (GOM) fishery for Atlantic cod(Gadus morhua), a mainstay for New England andCanadian fishermen (Rich 1929; O’Leary 1981), hasremained relatively productive and resilient to fishingpressure until recently, even though the stock biomasshas been declining for a number of years. In 1998,National Marine Fisheries Service (NMFS) reportedAtlantic cod stocks in the Gulf of Maine were overex-ploited and at extremely low biomass levels (Mayo etal. 1998). Historical studies describe the decline ofsome Atlantic cod stocks as long-term processes thatvary greatly in the short term and are related to humaninteraction (Hutchings and Meyers 1995). They implythat if such fisheries were to be restored, more effectivemanagement would have to be developed. Implicit toimproved management, however, is the need to iden-tify the population components of Atlantic cod found

in areas like the Gulf ofMaine and to betterunderstand how thosestructures relate to diver-sity within and amongspecies (Smedbol andStephenson 2001).

In the present article,I evaluated the distribu-tion and dynamics ofAtlantic cod in the Gulfof Maine (GOM) duringthe 1920s, a periodwhen cod were abun-dant (Rich 1929; Ames1997). Subpopulationsand spawning compo-nents were tentativelyidentified; their distribu-tion and spawning areaswere compared to recenttagging studies and codegg distribution surveys.The study area includedthe GOM lying north of

a line from northern Massachusetts Bay (42 30’N) toWrights Swell (42 30’N, 68W), thence northeast toYarmouth, N. S. (43 50’N, 66 7’W) (Figure 1).

A Review of Atlantic Cod Population Structure andDynamics in the Gulf of Maine. To clarify the termsused to describe population structure, the followingdefinitions were used: a population is a self-repro-ducing group of conspecific individuals that inhabitthe same range at the same time, are affected bysimilar environmental factors, and are reproduc-tively isolated from other populations. Asubpopulation is a semi-independent, self-repro-ducing group of individuals within a largerpopulation that undergoes some measurable butlimited exchange of individuals with other areaswithin a population. A spawning component is asegment of a population that does not differ ingenetics or growth, but occupies discrete spawningareas inter-annually. A stock is an arbitrary collec-tion of fish large enough to be essentiallyself-reproducing, with members of the unit exhibit-ing similar life history, and a local stock is a stockthat remains in a local area throughout the year(Smedbol and Stephenson 2001).

The GOM Atlantic cod stock is described asbeing one of three or possibly four groupings of codfound in northeast U. S. waters (Serchuk andWigley 1993). Groupings were reported to havelimited exchange with others (Wise 1963; Serchuket al. 1994), but the issue of genetic separationremains unclear. The GOM grouping was identifiedby length-frequency differences (Wise and Murray1957, 1958, 1959), parasite studies (Sherman andWise 1961), and meristic studies (Schmidt 1930),and was reported to contain many subdivisions(Wise 1963).

Reproduction, based on the maximum aver-age abundance of cod eggs in the Gulf of Maine,occurs in March (Berrien and Sibunka 1999). Inthe study area, however, coastal spawning areasexhibit a bimodal pattern with peak spawning inspring and fall. Spring spawning occurs in some

Atlantic Cod Stock Structure in the Gulf of Maine

Atlantic cod (Gadus morhua) in the Gulf of Maine provide an important but depletedfishery that needs to be made sustainable. However, restoring and maintaining robustpopulation components to achieve sustainability is made difficult when their distribu-tion and character is unknown. This study clarifies the structure of the Gulf of Mainecod grouping by deriving the distribution, movements, and behavior of populationcomponents from 1920s data and surveys of retired fishermen. These derivations areconsistent with current cod populations and with the existence of localized spawningcomponents. Nearly half the coastal spawning grounds of 50 to 70 years ago are aban-doned today and their spawning components have disappeared, suggesting depletion,undetected by system-wide assessments, may have been well advanced by the 1980s.

AB

STR

AC

T

Edward P. Ames Ames is a commercialfisherman with theStonington FisheriesAlliance and PenobscotEast Resource Center inStonington, Maine. Thisstudy is supported by theGulf of Maine ResearchInstitute. He can bereached at 207/367-2473or [email protected].

HU

NT

HO

WELL

Figure 1. A map of thestudy area shows thelocation of figures used inthe study.

fisheries researach

featu

reareas of the Bay of Fundy (Neilsen and Perley1996) and on the inner historical spawningground locations along the New England coast(Berrien and Sibunka 1999). Fall spawning gen-erally occurs at the outer spawning locations andin coastal areas of the Bay of Fundy (McKenzie1934). Recent studies indicate cod return tonatal spawning areas (Perkins et al. 1997;Wroblewski 1998; Green and Wroblewski 2000).Gulf of Maine Atlantic cod spawning groundsand juvenile habitats are on the coastal shelfand within approximately 37 km of the shore(Bigelow and Schroeder 1953; Ames 1997;Berrien and Sibunka 1999).

With the approach of spawning season codstart migrating towards their respective spawn-ing grounds, often appearing to movesequentially closer (Perkins et al. 1997).Spawning migrations close to the shore andrivers of the Gulf of Maine historically involvedlarge fish, but during the 1920s, the average sizeof cod on most other spawning grounds wassmaller (Bigelow and Schroeder 1953).

After arriving at their spawning area,Atlantic cod often gather into large schools(Earll 1880; Smedbol 1997). Being broadcast

spawners, older adults release small quantities ofeggs into the water column over a period of timeuntil spent. Depending on salinity, cod eggs usu-ally float to the surface and are graduallydispersed by winds and currents. Eggs hatch intwo or more weeks, depending on water temper-ature, and in another month or more, theresulting larvae metamorphose to juveniles andsettle to the bottom. Predation is very high onyoung juveniles during this phase unless theyquickly find shelter within interstices of theproper size among substrate particles (Gotceitasand Brown 1993).

Once spawning is completed, Atlantic codleave their spawning areas to pursue foragestocks in various feeding areas and initiate thenext annual cycle.

Methods

Sources of 1880s Fishing and Spawning GroundInformation: The 1880s database of cod groundsrelied heavily on descriptions and maps found inCollins and Rathbun (1887) and Earll (1880).Collins and Rathbun gathered their informationfrom interviews with fishing captains of the

January 2004 | www.fisheries.org | Fisheries 11

fishe

ries

rese

arch

featu

re


period, while Earll focused on cod spawningactivity in Ipswich Bay. The limited number ofappropriate records used to construct the 1880sdatabase restricted its use to evaluation of thepersistence of cod on particular fishing grounds.

Sources of 1920s Fishing and Spawning GroundInformation: The 1920s database formed thebasis for the current study. The cod fishinggrounds of Rich (1929), supplemented byBigelow and Schroeder (1953) and Ames(1997), provided historical fishing ground infor-mation. Rich gathered his data from directinterviews with “a large number of fishing cap-tains of long experience upon these grounds”and “in cases of conflict of their opinion, thegreatest agreement as to the facts has beenaccepted.” His interviews were conducted duringthe 1920s and included “a large number ofgrounds described earlier by G. Browne Goode”(Collins and Rathbun 1887). Most interviewswere with experienced captains using hook-and-line techniques. Additional grounds came fromAmes (1997), who collected fishers’ ecologicalknowledge (FEK) from 28 interviews withretired fishing captains, many of whom operatedotter trawlers from 1930-1960, and the classicwork by Bigelow and Schroeder (1953) thatsummarized 1920s and 1930s fisheries researchfrom U.S. Bureau of Commercial Fisheries.

Cod spawning areas of the period identified byAmes (1997) were supplemented by additionalgrounds from Bigelow and Schroeder (1953) andCoon (1998), who had gathered FEK in the Bayof Fundy area about where ripe Atlantic cod werecaught. Cod egg distribution studies byMcKenzie (1934) were used to confirm locationsidentified by Coon (1998). Cod spawning areacriteria (Ames 1997) required that spawningareas have two or more independent reports ofripe cod being caught on site during spawningseason, with depths of 10 m to 100 m, and thatthe substrate was sand, muddy sand, muddygravel, or a muddy basin grading to sand along itsedges. Much of this information came from ottertrawlers of the period. Spawning areas were oftendepressions of muddy gravel and sand borderingcod feeding grounds. Bigelow and Schroeder didnot discuss spawning area substrates, but NOAAcharts indicated the areas identified had appro-priate depths and substrates.

Gulf of Maine spawning seasons for cod wereobtained from historical reports (Earll 1880;McKenzie 1934; Bigelow and Schroeder 1953)and recent cod egg distribution surveys (Nielsenand Perley 1996; Berrien and Sibunka 1999).Cod spawning peaked between September andDecember in the GOM and continued at moder-ate levels from January through May. Severalareas had two or more spawning events in thesame year. Cod eggs continue to be reported at

many Gulf of Maine locations each year, butrecent spawning events east of Casco Bay havebeen intermittent and small (Berrien andSibunka 1999).

Locating Historical GOM Fishing and SpawningGrounds. Historical fishing and spawninggrounds were compiled in a GIS system by fol-lowing historical navigation instructions with anavigation program using digitized NOAA nau-tical charts (Ames 2001). Single or multiplepoint bearings were extrapolated to a point nearthe ground. From the immediate vicinity of thispoint, a location was selected that had theappropriate scale, orientation, and benthic char-acteristics consistent with that described byfishermen (Figure 2a). In the case of spawningareas, these characteristics were in agreementwith the depth and substrates of other knownspawning habitats of Atlantic cod, e.g., muddydepressions of sand or gravel or hard mud bot-tom surrounded by cobble or gravel. Thisprocedure imparts a visual precision to the fig-ures that, absent this clarification, would appearto overstate the information contained in thehistorical navigation instructions.

Determining Seasonal Distributions of AtlanticCod Using Relative Availability (RA). Maps wereprepared for each season of the year showing dis-tribution based on fishermen’s estimates ofAtlantic cod concentrations on cod grounds(Rich 1929). Fishermen normally estimated theavailability of cod by describing fishing as good,fair, poor, etc. It provided a convenient way toshare information with each other, regardless ofthe size of the vessel, or type of gear being used,and avoided the need to share vessel landingswhich were proprietary information and oftenunavailable. When placed in appropriate spatialand temporal context, their observations of codconcentrations allowed their knowledge to beapplied to tracking historical cod shifts in codavailability.

Relative availability (RA) quantified thoseestimates for each season on each specific fish-ing ground. It provides no actual measure oftonnage landed, but refers to which season fish-ermen deemed best, average, or poorest forcatching cod on a particular ground. High RAvalues occurred when large seasonal concentra-tions were present and described optimalconditions for cod to gather on a particular fish-ing ground. RA values ranging from 0-4 (Table1) were used to establish color gradients on GISdisplays of fishing grounds to visualize seasonalspatial distributions of cod availabilities (Figures3a,b,c,d).

Determining Historical Movements of AtlanticCod with RA values. In addition to determiningseasonal distribution, RA values were used totrack 1920s cod movements and migrations

fisheries research

featu

rebetween seasons. Grounds recording an increasein RA during a seasonal change were assumed tohave received additional cod from the nearestbordering fishing ground reporting a lower RAvalue. In other words, it was assumed that codminimize the distance traveled in migration.One could imagine river-like topographieswhere fish had to pass one site in order to reachothers, but, with some exceptions, this is not thecase in the GOM.

The direction of RA movement between bor-dering grounds was described on GIS with anarrow drawn between the two borderinggrounds. This procedure yielded unambiguousresults when tracking local movement patternsof isolated cod concentrations that linked thewinter locations of subpopulations with histori-cal spawning areas (Figure 4). However, theidentities were obscured when two or more codconcentrations overlapped. Cod movementscould only be tracked along the grounds

mapped, but cod were assumed to also inhabitareas surrounding the grounds at times and over-laps of movement were interpreted to be generalseasonal movements (Figures 5a,b).

Regions having a broad-based, continuousdirection of RA movement involving severalgrounds one season and accompanied by a simi-lar, opposite movement during a later seasonalchange, were tentatively classified as migrationcorridors (Figure 6). These areas showed migra-tion patterns similar to recent tagging studies ineastern GOM (Hunt et al. 1998) and westernGOM (Perkins et al. 1997).

Identifying Historical Spawning Components.Spawning components were identified by track-ing seasonal RA values through an annual cycle.A given cod concentration on a ground experi-encing a seasonal decrease in RA value wasassumed to have moved to the nearest groundshowing an increase in RA. Concentrationswere tracked only where cod were present all


Figure 2a. Historical fishing grounds and spawning areas of Atlantic cod in the Gulf of Maine were used to create an X-Y plot for trackingAtlantic cod movements.

fishe

ries

rese

arac

h

featu

re


year and their shifts were obvious. Cod concen-trations that were present all year anddemonstrated cyclic RA movement patternsbetween their winter grounds and a specific his-torical spawning area that was occupied duringspawning season were tentatively identified asspawning components. Many of these historicalspawning sites are still active while others areabandoned and their spawning componentshave disappeared (Ames 1997; Berrien andSibunka 1999).

These mapping procedures produce patternsthat are strongly consistent with the existence oflocal populations and were distributed along theentire length of the northern GOM coastline.

Experimental Design: Edge Effects. Seasonalmigrations and movements of cod concentra-tions were tracked on an X–Y grid of fishinggrounds. This strategy worked well as long as themovement remained within the grid. However,when fish movement occurred at the edge of thegrid (for instance, movements during the sum-mer in Penobscot Bay) they were arbitrarily

assumed to move in the same direction as othercod in the immediate area.

ResultsCharacterizing the Gulf of Maine’s Historical

Atlantic Cod Grounds. In all, approximately 260fishing grounds (Rich 1929; Ames 1997) and 91spawning grounds (Bigelow and Schroeder 1953;Ames 1997; Coon 1998) were distributedbetween Cape Ann in western GOM and theLurcher Shoal in eastern GOM (Figure 2a).Inshore cod feeding grounds were generally areasof rocky bottom along the 100-m depth contouron the coastal shelf and bordered deeper coastaldepressions and channels that were used asspawning areas in some locations. Depths on thegrounds ranged from a few meters to more than200 m. Atlantic cod grounds further offshoreincluded the banks, ridges, and swells that liebetween the major basins of the gulf. Most his-torical fishing grounds were located on the shelfwithin 45 km of land.

Coastal spawning grounds of Atlantic codwere either contiguous with, or when in the

vicinity of major bays, often closer to shore thanfishing grounds of the period. Spawningoccurred in channels and basins bordering therocky, shallow historical fishing groundsdescribed by Rich and generally had bottoms ofmuddy gravel, sand, or mud with borders ofgravel and varied in depth from less than 10 mto 100 m.

Identifying Essential Fish Habitats of AtlanticCod in the Gulf of Maine. The selectivity of ottertrawling and hook fishing, the two major fishingtechnologies of the 1920s, lend insight into theseasonality of different habitats by Atlantic cod.During the 1920s, Atlantic cod in the GOMwere most commonly caught on baited hooks, atechnology selective for foraging fish. Decades ofdirected fishing with hooks on the Atlanticcod’s historical fishing grounds had establishedthose grounds as preferred foraging habitats forAtlantic cod in the GOM at certain times ofyear. Even though the particular forage speciesthat once attracted cod to feed at these locationsremains unclear, the location of the cod’s feed-ing areas and the seasons they were presentallows thousands of square kilometers ofAtlantic cod habitat to be identified.

In similar fashion, captains of early ottertrawlers discovered that cod gathered along themargins of basins and channels bordering histor-ical fishing grounds when not feeding, and insome locations, used the channels and basins asspawning areas.

Determining the Long-term Productivity of Codon Historical Atlantic Cod Fishing Grounds. Ofthe 260 fishing grounds from the 1920s used inthe study (Rich 1929, Ames 1997), 92 werebeing fished prior to 1880 (Collins and Rathbun1887). All of the 1880s grounds continued to beproductive in the 1920s, though several hadreduced landings. By the 1980s, 64 of the 1880sgrounds were still producing cod, while 26 innergrounds of the 1880s reported no cod.

The safety and effectiveness of motor-driventechnologies rapidly displaced coastal fishingvessels relying on sail. Their development madeinner grounds very accessible in winter and bythe 1930s, they had become the favored fishinggrounds of two new fishing technologies, ottertrawlers and coastal gillnetters (Alexander 1914;Ames 1997). Their introduction, combined withthe cumulative depletion of anadromous foragestocks caused by dams and coastal industrializa-tion (Baird 1883), were major factors in thedemise of the coastal cod fishery.

Evaluation of 1920s Seasonal Distribution. Inspite of their wide distribution, Atlantic codwere present on most 1920s fishing grounds foronly a few months of the year. Of the 260grounds surveyed during the 1920s, only 60 wereyear-round fishing grounds (Figure 2b). All but

Table 1. Fishermens’ descriptions of cod fishing were standardized into relativeavailability (RA) values.

Descriptive comments Atlantic cod concentrationsNot mentioned or absent 0Poor fishing 1Fair or present with no estimate 2Good or abundant 3Excellent or very abundant, etc. 4

17 were deeper, offshore grounds with noreported spawning activity. Most of the remain-ing 200 grounds were cod grounds for twoseasons of the year or less.

Of the 88 grounds reporting cod in winterand 161 grounds reporting in spring, 54% hadvery good cod fishing. It seems likely that theGulf of Maine’s fall and spring spawning seasonsmay have influenced these results. Fishermenfound fewer grounds with good fishing (RA=3)in summer and fall (38% and 41% respectively),suggesting that Atlantic cod were more dis-persed after spawning. The total number of codfishing grounds and number of grounds withgood fishing were summarized for each season(Table 2).

Identifying the Seasonal Movements of 1920sAtlantic Cod. Migrations and local movementswere based on seasonal shifts of 1920s cod con-centrations among fishing grounds. Theseoccurred on both sides of the Gulf of Maine(Figure 6a) and agreed well with recent taggingstudies (Perkins et al. 1997; Hunt et al. 1998).To avoid confusion in the current article, move-ments refer to cyclic movement patternsbetween the winter grounds of one of four largeconcentrations of cod and historical spawningareas lying nearby while migrations refer tolarge-scale RA movement patterns that extendbeyond the winter grounds.

Winter, the Winter Buffer Zone, and theSeparation of Inshore and Offshore Populations. Inwinter a continuous band of fishing grounds,sandwiched between the coast and CashesLedge and extending from north of JeffreysLedge to Grand Manan Bank, was abandoned byAtlantic cod. This appears as a zone separatinginshore and offshore populations (Figure 3d).The 1920s buffer zone included more than 14cod fishing grounds that were productive formuch of the year, but had no cod landings inwinter. At the same time, good winter cod fish-ing was reported on either side of the bufferzone.

The 1920s winter buffer zone is significantbecause it provides evidence of a separationbetween coastal cod and offshore cod preciselyat the time of year when many Atlantic cod arefound close to their spawning grounds and sug-gests that fishing grounds in the winter bufferzone were seasonal feeding areas that had beenabandoned as spawning season approached.

Historical Atlantic Cod Migrations in the Gulfof Maine. When movement patterns betweenbordering fishing grounds in the GOM areviewed as a whole, the direction of historicalmigrations are evident. Offshore concentrationsof 1920s Atlantic cod in the GOM generallymigrated north in spring and south in fall, whileinner coastal shelf cod generally had inshore-off-shore movement patterns (Figure 6a).


Figure 2b. Year-roundAtlantic cod groundslocated offshore had nospawning activity.

fisheries research

featu

re

Movements and migrations among historical fishing groundswere in agreement with recent tagging studies (Perkins et al.1997; Hunt et al. 1998).

Discussion Determining Spatial Features of 1920s Atlantic Cod Stock

Structure: Gulf-wide overviews were constructed to track seasonalshifts in cod availability (RA) in an effort to evaluate fine-scalepopulation structures in the Gulf of Mainegrouping. The overviews revealed how vari-ous parts of the 1920s grouping functioned asa system when the stock was robust. Thisallowed the interactions of various parts ofthe system to be studied in detail.

Wise (1963), during his characteriza-tion of the Gulf of Maine grouping,detected smaller population componentsthat he described as subdivisions. The fourlarge winter concentrations of Atlanticcod distributed along the coastal shelf wereassumed to be those subdivisions inhabit-ing the study area. Further examinationshowed that some cod concentrations inthe subdivision migrated seasonally andreturned to the same spawning areas annu-ally, while others stayed nearby. Localshifts of non-migrating cod in each subdi-vision between the winter grounds andspecific local spawning grounds revealedseveral areas where cod behaved likespawning components and local stocks(Smedbol and Stephenson 2001).

Subpopulations: Each subdivision had dis-crete migration corridors, local cyclicmovements to and from spawning grounds,nursery areas, and was partially isolated bybathymetry. The four subdivisions identi-fied in the study were composed of clustersof spawning components or local stocks.That is, they functioned as subpopulations(Table 3).

This classification was further supportedby the area-specific nature of their depletion,the co-extinction of several 1920s coastalspawning components associated with theloss of anadromous forage stocks such asalewives, the demonstration of local, long-term reproductive capacity from correlationof 1920s spawning sites with 1977-1987 codegg distribution patterns, and validation of1920s historical migrations and movementsfrom two recent tagging studies in eastern andwestern Gulf of Maine.

Several features were common to Gulf ofMaine subpopulations. In winter, cod gath-ered into major concentrations (Figure 3d)via separate movement patterns (Figure 6b).Each subdivision occupied a relatively dis-tinct region on the coastal shelf that appeared

to be partially separated from the others by deeper basins. sub-populations had several spawning grounds (Figure 2a), nurseryareas, and a seasonal migration corridor abutting its wintergrounds. Migration corridors led away from the winter groundsfollowing local bathymetry and in some instances, involved theridges and swells separating the major basins (Figure 6).

Significant numbers of Atlantic cod also remain offshore dur-ing spawning season. They are found in large numbers all year onthe Cashes Ledge complex and the rises and swells rimming the


Figure 3a. The distribution of Atlantic cod and relative size of their concentrations in springis displayed as a color gradient of fishing grounds.

Figure 3b. The distribution of Atlantic cod and relative size of their concentrations as relativeavailability (RA) values in summer is displayed as a color gradient of fishing grounds.

Gulf’s major basins, making them very productive fishing grounds(Figure 2b). Rich (1929) described these areas as being migrationconduits and feeding areas for Atlantic cod. No historical reportsof ripe cod were found for the offshore areas and the 11-year codegg survey by Berrien and Sibunka (1999) found but a single iso-lated spawning event near the ridge south of Fippinies.Movement patterns support their function as feeding areas forreproductively resting cod from bordering subpopulations.

Spawning Components / Local Populations. Using a list of cri-teria from available 1920s data (Table 4), 16 historicalspawning components were tentatively identified among the 4subpopulations. Identification was limited to areas whereAtlantic cod concentrations remained in the subpopulation’sarea all year and seasonal shifts between winter fishing groundsand a spawning ground were obvious. This provided evidencethat Atlantic cod were using specific spawning areas in eachsubdivision. All are described as spawning components, even

though many spawning grounds in the east-ern GOM once active in spring are nowinactive and may have involved local popu-lations that are now extinct.

The territory and movement patterns oftwo typical historical spawning compo-nents from the Midcoast subpopulation areshown; an extinct, inner westernPenobscot Bay component that spawned inspring and disappeared in the 1940s, and arecently active spawning component ofouter eastern Penobscot Bay that spawnedin the fall (Figure 5).

Concentrations of eastern Penobscot BayAtlantic cod returned from their offshoresummer grounds in the fall to spawn onBowdies and Gravelly (south of MatinicusSeal Island) where they remained throughspring. Simultaneously, the inner westernPenobscot Bay spawning component movedvia Green Island Ridge, to grounds north ofMatinicus and Seal Island where theyremained until spawning in late winter/earlyspring. By summer, both components left thePenobscot Bay area for their summer groundswhere they remained until fall.

The GOM demonstrates a remarkablevariation in spawning times of Atlantic cod,with spawning occurring all year except mid-summer. From 1977-1987, the winter groundsof historical subpopulations were usuallyspawning sites in the fall (Berrien andSibunka 1999). Additional spawning compo-nents used inner spawning areas in spring.

The Interactions of Historical GOMSubpopulations and Spawning Components.Recent studies have given insight into wayshistorical subpopulations and spawning com-ponents interacted during seasonalmigrations. The western GOM subpopula-tion, for example, is the most robust of theAtlantic cod subpopulations (Figure 8) and isdominated by Ipswich Bay’s historical winterfishery for spawning Atlantic cod (Earll 1880;Bigelow and Schroeder 1953). With largeconcentrations of cod appearing in IpswichBay to spawn, it has always seemed logical toassume that other coastal spawning areaswere not significant to either the fisheries ofwestern GOM or Ipswich Bay (Bigelow andSchroeder 1953). Historical migration corri-


Figure 3c. The distribution of Atlantic cod and relative size of their concentrations as relativeavailability (RA) values in fall is displayed as a color gradient of fishing grounds.

Figure 3d. The distribution of Atlantic cod concentrations in winter and relative size of theirconcentrations is displayed as a color gradient of fishing grounds. The unoccupied cod grounds inthe Winter Buffer Zone are also identified.

fishe

ries

resd

arch

featu

re


dors appeared to be simple transport corridors mov-ing pre-spawning and spent fish to and from thetwo major spawning areas. However, that explana-tion no longer seems adequate.

Atlantic cod migrations to Ipswich Bay appearto be much more than simple spawning migrations.In fall, Atlantic cod tagged in the Sheepscot Bayregion (Perkins et al. 1997) were shown to migrateprogressively closer to Ipswich Bay, where theywere caught during winter. They subsequently leftIpswich Bay in spring and returned to theSheepscot area. The Perkins study showed thatSheepscot Bay cod contributed substantially toIpswich Bay’s winter cod fishery, though thereverse has not been established. This would havemade Ipswich Bay landings in winter appear mis-leadingly large.

Berrien and Sibunka (1999) reported thatspawning events in the Sheepscot area were com-mon in May and November and in Ipswich Bay, inMarch-April, confirming historical observations(Earll 1880; Bigelow and Schroeder 1953). Spring

spawning occurs in Sheepscot Bay spawning areasas Atlantic cod arrive from Ipswich Bay (Perkins etal. 1997). Spawning also occurs in late fall on outerSheepscot spawning areas after the spring-spawn-ing cod have migrated, and indicates two spawningcomponents coexisted in the Sheepscot area.

By spring, Ipswich Bay cod had finishedspawning and were leaving the Ipswich area, dis-persing northward as part of a feeding migration(Bigelow and Schroeder 1953). Yet, SheepscotBay cod returning from Ipswich Bay were on theirspring spawning migration to Sheepscot Bay(Perkins et al. 1997). The migration pattern isexactly opposite that of Ipswich Bay cod anddescribes the mixing of spawning components inthe western subpopulation.

However, their seasonal migrations coincidewith seasonal migrations of Atlantic herring(Clupea harengus) (Rich 1929; Bigelow andSchroeder 1953), raising the possibility that thepursuit of migrating forage stocks are involved inthe reverse migration of Sheepscot Bay. Cod are

Figure 4. Sequential movements of Atlantic cod concentrations in the Wells Bay area are shown as they move from summer fishing grounds tothose used in fall.

fisheries research

featu

reone of the three major predators of Atlantic herring(Collete and MacPhee 2002) and during the 1920s,both species overwintered in Ipswich Bay (Rich1929; Bigelow and Schroeder 1953). Both spawningcomponents migrated with herring in spring andfall, resulting in one spawning component being ona spawning migration while the other was on a feed-ing migration.

This would not be unusual. Coastal Atlanticcod also used to co-migrate in spring with alewives(Alosa pseudoharengus) and blueback herring(Alosa aestivalis) as they returned to spawn in natalrivers and streams (Baird 1883). Baird reportedthat the loss of these forage stocks had triggeredthe collapse of the coastal cod fishery (Baird 1883)and the abandoning of spawning grounds lyingclose to rivers.

Such a pattern suggests that Atlantic herringand similar forage stocks may provide the impetusfor Atlantic cod migrations and opens the possibil-ity that cod are “programmed” to arrive at theirspawning areas via their pursuit of a particular for-

age stock sharing a common migration corridor.The disappearance of local anadromous foragestocks and the disappearance of nearby Atlanticcod spawning components was a coincidence thatoccurred in several areas (Figure 9), suggesting thatthe traditional movement patterns and arrival timesof Atlantic cod may have been disrupted at theirinner spawning grounds when the forage stock dis-appeared. If so, the restoration of coastalpopulations of Atlantic herring, alewives, and riverherring may also be important to restoration ofcoastal Atlantic cod fisheries.

Further Issues

While the study has demonstrated a relation-ship between historical and recent stockcomponents, several issues challenge the rela-tionship’s validity and must be addressed.

Should inferred movements of cod stocks be usedto make conclusions about stock structure? Thestudy was based on the hypothesis that if Gulf ofMaine cod concentrations could be tracked


Figure 5. Local movement patterns of Atlantic cod of the two Penobscot Bay spawning components moved along different corridors, occupieddifferent spawning grounds, and spawned at different times of year. Cashes Ledge lies slightly south (bottom) of the figure.

A B

fishe

ries

rese

arch

featu

re


through the year, the location of populationcomponents in the grouping could be identifiedfrom their behavior and movements. Using theAtlantic cod’s tendency to home to specificspawning areas, spawning grounds of the periodwere used as points of origin for identifying andtracking cod concentrations.

The protocol for tracking concentrations ofAtlantic cod infers that they move conserva-tively; that is, they move from one fishing groundto the nearest bordering fishing ground offeringappropriate conditions and habitat. Thesesequential movements along fishing groundshave long been recognized and exploited by NewEngland fishermen, particularly in western Gulfof Maine. In fact, the inferred historical move-ments agreed with the results of the Perkinstagging study (1997) and the Hunt tagging study(1998) confirming the methodology used toidentify movement patterns.

Should subdivisions of Atlantic cod in the Gulf ofMaine grouping be described as subpopulations?The four large concentrations of 1920s cod dis-tributed in different sections of the study areawere identified as the subdivisions mentioned inWise’s characterization of the Gulf of Mainegrouping (1963). Depletion patterns indicatedthat the number of cod in a subdivision did notvary with reductions in other subdivisions andbathymetric charts show the subdivisions to bepartially isolated from each other by deeperbasins. Each area contained spawning grounds,nursery areas, and separate migration corridors.These features are not characteristic of migrat-ing, pandemic populations, but are quite typicalof discrete population units. They are, in fact,characteristic of semi-independent, self-repro-ducing groups within a larger population thatundergo limited exchanges within a population(Smedbol and Stephenson 2001).

Closer examination revealed the subdivisionswere composed of several bodies of cod thatdemonstrated separate circular movement pat-terns linking local winter grounds to specificlocal spawning areas; that is, they functioned asan assemblage of spawning components usingseparate spawning grounds in the area of a givensubdivision. Berrien and Sibunka (1999) havereported recent spawning activity on many ofthese areas, indicating their continued activity.While it may be that subdivisions are but singlespawning components among one or more localstocks, their behavior within the grouping isbest describes as that of subpopulations(Smedbol and Stephenson 2001). In theabsence of objective studies that differentiatebetween bordering spawning components, theevidence seems persuasive that subdivisions aresubpopulations of the Gulf of Maine grouping.

Accuracy of charts and navigation techniques ofthe period. Much of the data used in the studypredated electronic navigation. Prior to itsdevelopment, fishermen located fishing groundsusing sextant or compass bearings and distancesfrom known landmarks. They were able to comeclose to a given fishing ground using eithermethod, but to confirm their location fishermenhad to sample the substrate and depth in thearea with a sounding lead until they found theright spot. Grease stuck on the bottom of thelead provided them with a sample of the oceanbottom around the fishing ground.

Period fishermen repeatedly returned to thesame fishing ground by following the same nav-igation instructions and correlating bottomcharacteristics with known bathymetric descrip-tions of the ground. While most fishing groundsidentified in the study could be readily found,the precise location of some grounds was limitedby the accuracy of bathymetric information onmodern NOAA charts.

Winter Spring Summer Fall Year-round# Grounds with cod: 88 161 117 134 58Good fishing or better: 54% 54% 38% 41% 22%

1. Winter grounds were proximal to coastal spawning areas. 2. Spawning areas bordered deeper channels or basins.3. Atlantic cod concentrations were present in the general area all year.4. Cod concentrations could be tracked between a specific historical spawning

ground and local winter grounds during spawning season. 5. Ripe Atlantic cod were confirmed to have been present in the spawning area

during spawning season by two or more independent observers.6. Spawning grounds had appropriate depths and substrates.

* The above criteria were derived from parameters available from 1920s data and are not intended toreplace today’s more rigorous criteria defining cod populations.

1. The Western subpopulation occupied the coastal shelf from northernMassachusetts Bay and Ipswich Bay to the vicinity of Sheepscot Bay andincluded six spawning components.

2. The Midcoast subpopulation in the Penobscot Bay area occupied the coastalshelf from Muscongus Bay to Jericho Bay and included four spawningcomponents.

3. The Downeast subpopulation in the Mt. Desert Island area had threespawning components distributed from Mt. Desert Rock to Petit Manan.

4. The Bay of Fundy sub-population had three spawning components andincluded cod from Passamaquoddy Bay to the WNW Rips and possiblyGerman Bank.

* Similar behavior was noted outside the study area, particularly in Massachusetts Bay. However, seasonalmovement patterns in Massachusetts Bay showed that cod concentrations moved east and south ofStellwagen Bank and outside the study area. Only two small spawning areas along its northern edgeappeared to have cod that moved northeast in spring and returned from the north in fall.

Table 2. The number of 1920 fishing grounds with landings of Atlantic cod.

Table 3. Gulf of Maine Subpopulations of Atlantic cod*.

Table 4. Criteria for identifying historical spawning components of Atlantic cod in the GOM*.

fisheries research

featu

re


Figure 6a. Historical migration patterns of Atlantic cod in the Gulf of Maine followed contours of the coastal shelf and deeper offshore ridges andshoals. Recent tagging studies (circled areas) agreed with historical movements. Figure 6b. Historical spawning grounds (circled areas) were at the terminus of migration corridors used by specific sub-populations of Atlantic cod.

fishe

ries

rese

arch

featu

re


Bias and Historical Fishery Information. Differentfishing methodologies generate different types ofinformation, making it important to evaluate datacarefully. For example, the prevailing technologyprior to 1930 was still hook and line in coastal Gulfof Maine, which caught only foraging Atlantic cod.When fish refused to bite baited hooks, fishermencould only assume there were no cod in the area, forthey had no other way to confirm their absence.However, this unintentional bias has provided aconvenient way to locate the major forage habitatsof Atlantic cod and identify the times of yearAtlantic cod were foraging on each site.

By contrast, information about spawningAtlantic cod, which are less inclined to feed, gen-erally came from fishing methods not relying onfeeding behavior. Predictably, most spawningground information came from interviews withinshore otter trawler fishermen of the 1930s, 1940s,and 1950s that had once targeted coastal cod

spawning aggregations. The advent of the ottertrawl was the first time cod spawning aggregationshad been vulnerable to fishing gear on such a largescale. Early otter trawlers generally towed their netson smoother bottom than that used by hook fisher-men and because they were mobile, often could nottell exactly where they caught fish, even thoughcleaning the catch allowed them to identify whenfish were feeding, fasting, or spawning. The rapidcollapse of coastal cod stocks exposed to ottertrawling may have demonstrated the great suscepti-bility of inshore Atlantic cod to that technology(Ames et al. 2000).

Migration patterns of the 1920s may have beentransient events. Fishermen often refer to migrationsand local movements of Atlantic cod and whiletagging studies have clarified cod movements insome areas (Perkins et al. 1997; Hunt et al. 1998),the persistence and fine-scale details of cod move-ments in the Gulf of Maine have been poorly

Figure 7. Recent cod egg surveys confirm that many spawning components of historical Gulf of Maine subpopulations are still active.

fisheries research

featu

reunderstood. It is possible that historical movementpatterns may only be valid for brief periods, makinga comparison with recent movement patternsunjustified. To test this hypothesis, 1920s seasonalmigration patterns of Atlantic cod (Figure 6) werecompared to the results of recent tagging studies(Perkins et al. 1997; Hunt and Neilson 1993; Huntet al. 1998).

The comparison showed that historical migra-tions of Atlantic cod occurred in the same areas,during the same seasons, and with fish moving inthe same direction as in recent migrations. Thetest confirmed that the same migration routeshave been followed for more than a century andreflect long-term responses of Atlantic cod tounderlying ecological factors. This refutes theargument that Atlantic cod migrations were tran-sient events that varied significantly in time andplace. To the point, historical movement pat-terns may be useful in identifying the movementsof today’s Atlantic cod populations when theyoccur in the same area.

Concern exists that coastal spawning componentsof Atlantic cod have disappeared and may be extinct,making historical comparisons moot. If the histori-cal coastal stocks of Atlantic cod no longer exist,comparisons of their behavior patterns with thebehavior found in today’s spawning componentswould be irrelevant.

To test this, the following extinction hypoth-esis was developed: because spawningcomponents of subpopulations return to the same

spawning ground each year to reproduce, no ripefish or early stage cod eggs would be foundaround historical spawning areas where they wereextinct. If they still exist, ripe adults and cod eggswould be found on the spawning areas duringspawning season. By correlating historical spawn-ing grounds and/or pre-spawning aggregationswith recent cod egg surveys, grounds having highdensities of cod eggs would indicate a continuedpresence of Atlantic cod populations and alsogive supporting evidence that homing byAtlantic cod occurs in the Gulf of Maine.

Recent GOM cod egg distribution patterns(Berrien and Sibunka 1999) were plotted on GISand their locations were compared with the loca-tions of historical Atlantic cod spawninggrounds (Figure 7). The comparison showed that1980s GOM areas with high-densities of codeggs in the fall were the historical subpopulationwinter grounds. In spring, several historicalspawning grounds also were active, confirmingthe presence of continued spawning activity bysome historical spawning components. However,most spawning events were small and infrequent,indicating their depleted condition.

Approximately 40 of the 90 historical spawn-ing grounds had neither commercial landingsnor evidence of spawning during the 11-yearstudy of Berrien and Sibunka (1999) and appearto be extinct. This represents a significant reduc-tion in the GOM’s reproductive capacity andcoincides with a gradual, long-term depletion


Figure 8. Displaying codegg distribution andhistorical spawninggrounds identifies thelocation of activespawning components inthe Westernsubpopulation.

fishe

ries

rese

arch

featu

re


marked by large fluctuations in landings of thefishery, particularly in eastern GOM.

Depletion and recovery patterns of spawning com-ponents. By the late 1940s, coastal subpopulationswere depleted or collapsed (Maine Dept. Sea andShore 1947). Most of the spawning grounds aban-doned during this period were found in coastalwaters between Casco Bay and the Bay of Fundy.Many 1920s spawning components characterizedin the present study were associated with theseinactive spawning grounds (Figure 9). Activespawning sites are defined as those that are spa-tially coincidental with the egg distributionstudies of Berrien and Sibunka (1999).

Spatial complexity in Atlantic cod, the distri-bution of spawning components, has beenhypothesized to be a function of oceanographicprocesses during egg and larval stages (Sinclair1988) and results in several different populationswith separate spawning areas and discrete eggand larval distributions. In a review of within-

species diversity, Smedbol and Stephenson(2001) observed that cod and herring have com-plex population structures in the NorthwestAtlantic with multiple subpopulations and thatmanagers have often failed to prevent the loss ofspawning components in these heavily exploitedfisheries. They concluded that spatial populationstructure and dynamics may be important to themaintenance of such fisheries and recommendedthat fine-scale population structure be preserveduntil “the weight of evidence suggests that it isnot of ecological significance.” Frank et al.(1994) attributed the collapse of the Sable-Western Bank subpopulation to a dramaticincrease in exploitation and concluded the tar-geting of spawning aggregations resulted in lostreproductive capacity that led to the subpopula-tion’s collapse.

At least three human-induced factors con-tributed to the collapse of Atlantic cod spawningcomponents in the GOM—the increased effort

Figure 9. The absence of cod eggs on historical spawning grounds reveal the location of abandoned spawning grounds and lost spawningcomponents. Circled areas identify inactive spawning areas.

from the introduction of otter trawling and gillnetting in coastalwaters, the pollution of coastal nursery grounds, and the destruc-tion of anadromous forage stocks by the construction of dams.

Foremost among these factors may have been the targeting ofAtlantic cod spawning aggregations as theygathered in coastal basins and channels in win-ter by coastal otter trawlers and gillnetters aftertheir introduction during the 1930s and 1940s.Few fishermen of the period are still alive, butthe brief fishing bonanzas described in differentcoastal areas were predictably similar (Ames1997). More recent collapses of subpopulationswere concurrent with refinements in fishingelectronics and technology that allowed greaterexploitation of bottom habitats.

A second factor may have been the widespreaddegradation of coastal nursery grounds and estuar-ies by industrial pollution from rivers andstreams in areas such as Penobscot Bay (Ames1997). Though restoration efforts haveimproved water quality, cod and other commer-cial species have not repopulated theirabandoned areas, leaving open the possibilitythat extinct spawning components may have beendiscrete populations adapted to those localities.

Forage stocks suchas alewives and blue-backs were lost whendams were built topower New England’sfactories, and causedthe cod that pursuedthem to the mouths ofrivers and streams to disappear, triggering the col-lapse of the coastal cod fishery (Baird 1883).

The combination of lost forage stocks,degradation of nursery habitats, and directedfishing on coastal spawning aggregations ofAtlantic cod appears to have overwhelmed sub-populations and caused several coastalspawning components to disappear.Maintenance of remaining Atlantic cod spawn-ing components, spawning areas, and nurserygrounds seems critical to recovery if a robust,sustainable fishery is to be re-established.

Management Perspectives. In spite ofpeaks in productivity, the gradual declineand collapse of the coastal cod fishery inmid-coast and eastern GOM has been con-current with the disappearance of inshorespawning components and the extinction ofhistorical coastal spawning grounds. Activespawning grounds in the study area fell from90 in the 1920s to no more than 46 in the1980s, with many of the remaining groundsexhibiting only sporadic activity (Figure 9). Most lossesoccurred in the three eastern subpopulations.

Similar losses in reproductive capacity from collapsed spawn-ing components were noted in the collapse of the heavilyexploited Sable/Western Bank subpopulation (Frank 1994).

However, unlike the Sable/Western Bank subpopulation, deple-tion of coastal GOM escaped detection because system-wideassessments, the basis for current management strategies, cannotdetect the gradual erosion of spawning components. If, for exam-

ple, one assumes cod abundance to beequivalent among the four subpopulations,each would produce about 25% of the annualGOM landings. The collapse of a single sub-population would cause a comparablereduction in landings. Since the confidencelimits of typical groundfish assessments arealso in the range of 25%, it is probable thatthe collapse of a single subpopulation wouldbe undetected. However, if two or more sub-populations collapsed, annual landings would

be reduced by more than half (Midcoast, Eastern,and Bay of Fundy subpopulations were recentlydepleted).

This simple calculation identifies the geo-graphical character of subpopulations andspawning components as a pivotal factor inrebuilding and maintaining GOM stocks at high,sustainable levels. It also points out the conse-quences on a fishery when management reliesprimarily on system-wide assessments. Smedbol

and Stephenson (2001)have suggested thatmanaging subpopula-tions and spawningcomponents on the basisof their geographicalcharacter would addresssuch losses. A sound

approach perhaps lies in the direction proposedby Wilson (1997), who stated that decentralized,hierarchical management units equivalent to thescale of the Atlantic cod’s population structurewould be more effective. He concluded that (1)organizing a decentralized fisheries managementsystem and (2) creating individual incentivesthat are consistent with the goal of sustainabilityis best achieved through local authority over eco-logical events whose impacts are strictly local.The present study suggests that subpopulationsare an appropriate management unit to minimize

further losses of spawning components whilerebuilding the fishery.

AcknowledgementsAppreciation is extended to Jim Wilson,

professor, Resource Economics, University ofMaine; Linda Mercer, lab director, MaineDepartment of Marine Resources, SteveMurawski, chief, Stock Assessment Divisionand Charles Byrne, fisheries biologist,

National Marine Fisheries Service, for their encouragement andcontributions. Thanks also to constructive suggestions fromRobert Stephenson, Joe Wroblewski, and Steven Cadrin. Theproject was funded by the Davis Conservation Foundation,Island Foundation, and Maine Community Foundation.


A joint collaborative cod tagging study coordinated by the Gulf ofMaine Research Institute and involving Canadian and Americanfishermen, Canada's Department of Fisheries, and National MarineFisheries Service is currently under way. Tag returns in the nextseveral years are expected to reveal more details about codpopulation structure and movement patterns.

MA

RIA

REC

CH

IAM

AR

IA R

ECC

HIA

RO

DN

EY R

OU

ND

TR

EEH

UN

T H

OW

ELL


References

Alexander, A. B., H. F. Moore, and W. C. Kendall. 1914. Ottertrawl fishery. Appendix VI to the report to the U.S.Commissioner of Fisheries for 1914. Dept. of Commerce,Bureau of Commercial Fisheries, Washington, D. C.

Ames, E. P. 1997. Cod and haddock spawning grounds of the Gulfof Maine from Grand Manan to Ipswich Bay. Pages 55-64 inThe implications of localized fishery stocks. Natural Resource,Agriculture, and Engineering Service, Ithaca, NY.

_____. 2001; Locating the historical fishing grounds of the Gulf ofMaine to track cod movements with GIS. Proceedings from the2nd Biennial Coastal Geotools Conference, Charleston, SC,January 8-11, 2001. U. S. National Oceanic and AtmosphericAdministration. NOAA/CSC/20114-CD. CD-ROM. NOAACoastal Services Center, Charleston, SC.

Ames, E. P., S. Watson, and J. Wilson. 2000. Rethinkingoverfishing: insights from oral histories of retiredgroundfishermen. Pages 153-164 in B. Neis and L. Felt, ed.Finding our sea legs. Institute of Social and EconomicResearch., St. Johns, Newfoundland.

Baird, S. 1883. U.S. Commissioner of Fish and Fisheries Report of1883, pages xi-xiv. Washington, DC.

Berrien, P., and J. Sibunka. 1999. Distribution patterns of fisheggs in the U.S. NE Continental Shelf ecosystem. 1977-1987;NOAA TR NMFS 145.

Bigelow, H. B., and W. C. Schroeder. 1953. Fisheries Bulletin:Fishes of the Gulf of Maine. U. S. Fish and Wildlife Service,Washington, D.C.

Collette, B. B., and G. Klein-MacPhee. 2002. Bigelow andSchroeder’s fishes of the Gulf of Maine. Third ed. SmithsonianInstitution Press, Washington, D. C.

Collins, J. R., and W. C. Rathbun. 1887. Fishing grounds of theeastern coast of North America. In G. B. Goode, ed. Thefisheries and fishing industries of the U.S. U.S. GovernmentPrinting Office, Washington, D. C.

Coon, D. 1998. An ecological sketch of some Fundy fisheries.Conservation Council of New Brunswick, Canada.

Earll, R. E. 1880. A report on the history and present conditionof the shore cod fisheries of Cape Ann, Mass., together withnotes on the artificial propagation of the species. Pages 695-704; US Commission of Fish and Fisheries, Washington, D.C.

Frank, K. T., K. F. Drinkwater, and F. H. Page. 1994. Possiblecauses of recent trends and fluctuations in Scotian Shelf/Gulfof Maine cod stocks. International Commission for theExploration of the Seas (ICES) Marine Symposia 198:110-120.

Gotceitas, V., and J. Brown. 1993. Substrate selection by juvenileAtlantic cod (Gadus morhua): effects of predation. Ocealogia(1993) 93:31-37.

Green, J. M. and J. S. Wroblewski. 2000. Movement patterns ofAtlantic cod in Gilbert Bay, Labrador: evidence for bayresidency and spawning site fidelity. Journal of the MarineBiological Association of the United Kingdom 80:1077-1085.

Hunt, J. J and J. D. Nielson. 1993. Is there a separate stock ofAtlantic Cod in the western side of the Bay of Fundy? NorthAmerican Journal of Fisheries Management 13:421-436.

Hunt, J. J., W. T. Stobo, and F. Almeida. 1998. Movements ofAtlantic cod, Gadus morhua, tagged in the GOM area. FisheriesBulletin 97:842-886.

Hutchings, J. A. and R. A. Meyers. 1995. The biologicalcollapse of Atlantic Cod off Newfoundland and Labrador. Vol.3, An Island Living Series. Institute of Island Studies,Charlotte, P.E.I.

Maine Department of Sea and Shore Fisheries. 1947. AnnualReport. Maine Department of Sea and Shore Fisheries.Augusta.

Mayo, R. K., L. O. O’Brien, and S. E. Wigley. 1998. Assessmentof the Gulf of Maine Atlantic cod stock for 1998. NEFSC Ref.Doc. 98-13.

McKenzie, R. A. 1934. Progress report no. 13 of AtlanticBiological Station, St. Andrew’s, N.S. Biological Board ofCanada, St. Andrews, New Brunswick.

Neilsen, J. D. and Perley, P. 1996. Can ichthyoplankton data beused to describe spawning areas of marine fish? Pages 147-156in Report of the second workshop on Scotia-Fundy Groundfish.Canadian Technical Report of Fisheries and Aquatic Sciences2100.

O’Leary, W. M. 1981. Maine seafisheries, 1830-1890. Thesis,University of Maine, Orono.

Perkins, H. C., S. B. Chenoweth, and R. W. Langton. 1997.The Gulf of Maine Atlantic cod complex, patterns ofdistribution and movement of the Sheepscot Bay substock.Bulletin of the National Research Institute of Aquaculture,Supplement 3:101-107.

Rich, W. H. 1929. Fishing grounds of the Gulf of Maine. U.S.Commissioner of Fisheries, Washington, D.C.



Schmidt, J. 1930. Racial investigations. X. The Atlantic cod(Gadus callarias L.) and local races of the same. Comptes-Rendus des Travaux du Laboratoire Carlsberg 18(6).

Serchuk, F. M., and S. E. Wigley. 1993. Assessment andmanagement of the Georges Bank cod fishery: an historicalreview and evaluation. Journal of Northwest Atlantic FisheriesScience 13: 25-52.

Serchuk, F .M., Grosslein, M. D., Lough, R. G., Mountain, D.G., and O’Brien, L. 1994. Fishery and environmental factors

affecting trends and fluctuations in the Georges Bank and Gulfof Maine Atlantic cod stocks: an overview. ICES Mar. Sci.Symp. 198, 77-109.

Sherman, K., and J. P. Wise. 1961. Incidence of the cod parasiteLernaeocera branchialis L. in the New England area, and itspossible use as an indicator of cod populations. Limnology andOceanography 6(1):61-67.

Sinclair, M. 1988. Marine populations: an essay on populationregulation and speciation. University of Washington Press,Seattle, WA.

Smedbol, R. K. 1997. Management considerations in dealing withthe existence of localized stocks. In The implications oflocalized fishery stocks. Natural Resource, Agriculture, andEngineering Service, Ithaca, NY.

Smedbol, R. K., and R. Stephenson. 2001. The importance ofmanaging within-species diversity in cod and herring fisheriesof the north-western Atlantic. Journal of Fish Biology 59(Supplement A):109-128.

Wilson, James. 1997. The social and management implications oflocal stocks. Pages 189-198 in The implications of localized fishstocks. Natural Resource, Agriculture, and EngineeringService, Ithaca, NY.

Wise, J. P. 1963. Cod groups in the New England area. FisheryBulletin 63:1.

Wise, J. P., and H. E. Murray. 1957. United States-1957. Lengthfrequencies-cod. Sampling Yearbook 2:126. InternationalCommission for the Northwest Atlantic Fisheries, Halifax,Nova Scotia.

_____. 1958. United States-1958. Length frequencies-cod.Sampling Yearbook 3:70-71. International Commission for theNorthwest Atlantic Fisheries, Halifax, Nova Scotia.

_____. 1959. United States-1959. Length frequencies-cod.Sampling Yearbook 4:35-36. International Commission for theNorthwest Atlantic Fisheries, Halifax, Nova Scotia.

Wroblewski, J. S. 1998. Substocks of northern cod and localizedfisheries in Trinity Bay, Eastern Newfoundland and in GilbertBay, Southern Labrador. Pages 104-116 in The implications oflocalized fish stocks. Natural Resource, Agriculture, andEngineering Service, Ithaca, NY.

atlantic cod stock structure feature in the gulf of maine€¦ · national marine fisheries service...

Documents