ICES mar. Sei. Symp., 198: 510-519. 1994 Vertebrae numbers as an indicator for the recruitment mechanism of coastal cod of northern Norway Solveig Løken, Torstein Pedersen, and Erik Berg Løken, S., Pedersen, T ., and Berg, E. 1994. Vertebrae numbers as an indicator for the recruitment mechanism of coastal cod of northern Norway. - ICES mar. Sei. Symp., 198: 510-519. Cod (Gadus morhua L.) in northern Norway are divided into two main groups (coastal cod and Northeast Arctic cod), but both groups are managed as a single stock. The coastal cod inhabit coastal and fjord regions and differ from Northeast Arctic cod with respect to growth pattern, age and size at maturity, and migration and spawning areas. Cod from the coastal and Northeast Arctic groups have been distinguished on the basis of meristic characters (average number of vertebrae), otolith type, and habitat selection during different life stages. We investigated whether the coastal cod are self­ recruiting, or whether this group also contains vagrants from the larger group of Northeast Arctic cod. Cod juveniles in coastal areas segregate vertically during the settling process, and the juveniles that settle in shallow-water sites have morphometric characteristics similar to those of adult coastal cod. The effects of topography and depth that influence separation are discussed. It is concluded that coastal and Northeast Arctic cod recruits have different early life histories. Solveig Løken, Torstein Pedersen, and Erik Berg: Norwegian College o f Fishery Science, University o f Tromsø, N-9037 Tromsø, Norway. Introduction CC also differ from NA with respect to growth pattern and age and size at sexual maturity (Rollefsen, 1934, Cod (Gadus morhua L.) in northern Norway are divided 1954; Hylen, 1964a). Jakobsen (1987) points out that the into two main groups: Northeast Arctic cod (NA) and crucial question of whether CC and NA should be coastal cod (CC) (Rollefsen, 1934, 1954). CC inhabit regarded as separate stocks is dependent upon “whether coastal and fjord areas and undertake only short-range fry from some of the main spawning grounds of North­ migrations. The larger NA stock, however, migrates east Arctic cod drift to nursery areas on the coast or in from the feeding areas in the Barents Sea and near the fjords and acquire the characteristics of coastal cod Svalbard to spawning areas along the Norwegian coast from the environment”. In order to identify whether (Hylen, 1964a, b; Godø, 1986; Jakobsen, 1987; Berg­ vagrant juveniles originating from NA spawning areas stad et al., 1987). CC spawn in several fjords along the inhabit the CC habitat in fjords and along the coast, an coast (Rollefsen, 1954; Jakobsen, 1987), and large num­ identifying marker for NA juveniles is needed. bers of them spawn in the most important NA spawning Meristic characters have frequently been used in in­ area in Lofoten-Vestfjorden (Hylen, 1964a). The larval vestigations of the structure of cod stocks in the western and juvenile drift from the NA spawning grounds on the Atlantic (Templeman, 1981, 1983; Lear et al., 1981; Norwegian coast towards the NA juvenile habitat in the Lear and Wells, 1984). Vertebrae number in fish is fixed Barents Sea and near Svalbard have been well described during the early part of the embryonic period (Fahy, (Sundby et al., 1989). In contrast, there is very little 1976), and differences in the average number of ver­ information on larval and juvenile drift from the CC tebrae (AV) between the groups of cod have been spawning in fjords and it is uncertain whether the CC attributed to differences in temperature during the em­ group is mainly self-recruited or recruited by vagrants bryonic period (Dannevig, 1947; Brander, 1979; Lear et from the NA spawning. The separation of cod into CC al., 1981). In support of this. Brander (1979) found a and NA is based on differences in the sizes and shapes of negative correlation between field data on temperature otolith zones (Rollefsen, 1934), which are influenced by during the spawning period and AV in cod stocks. environmental factors (Godø and Moksness, 1987). The Investigations carried out in northern Norway during ICES mar. Sei. Symp.. 198(1994) Vertebrae numbers as an indicator 511 the 1920s and 1930s indicate that NA cod have higher Bottom-settled cod juveniles were caught in Malan­ AV than CC (Schmidt, 1930; Rollefsen, 1934, 1954). gen, Sørfjorden, and other localities around Tromsø in The biological significance of the differences in AV September and October 1992 (Fig. 1). Bottom trawling found in adult NA and CC are uncertain. The aim of this was used to obtain samples from deep water (>55 m), study is to investigate whether AV could be used as a and beach seining was employed in shallow water (0-6 natural mark to trace cod juveniles settling to the bottom m). Cod juveniles collected by beach seine in Finnmark in the CC habitat. After settlement, the group identity during 1992 were obtained from Finnmarksforskning, (CC or NA) could be determined by the otolith struc­ Hammerfest, and six samples taken by Danish seine on ture. Vertebrae numbers were measured to assess the the coast of Finnmark in 1990 were provided by the variability in vertebral numbers of CC and NA in their Institute of Marine Research in Bergen. feeding and spawning areas. Young cod (age 1 + ) were sampled from various locations ranging from the inner parts of fjords to the outer coast of Nordland, Troms, and Finnmark in 1991— Materials and methods 1993 (Fig. 1). Samples were taken with bottom trawling, and with Danish seine, at various depths. Spawning cod were collected on the spawning grounds in In the laboratory, fish length, weight, and sex were Lofoten and Malangen in March and April during the recorded and then flesh was cleaned off the bones to years 1991-1993 (Fig. 1). Cod were fished with a pelagic facilitate counting the number of vertebrae. 0-group trawl at a depth of 70 m in Lofoten, and using a bottom cod were examined under a binocular microscope. All trawl in Malangen. vertebral counts include the urostylar half-vertebrae. Pelagic juveniles collected from the Norwegian coast Otoliths were used for age determination and for charac­ and Barents Sea during July and August/September in terizing the fish as either coastal cod or Northeast Arctic 1991-1992 were obtained from the Institute of Marine cod based on the growth pattern criteria described by Research, Bergen. Juveniles were also collected at these Rollefsen (1933, 1934). If available, 30 fish were ana­ times at locations along the coast and in the fjords of lysed in each sample of 0-group fish. Troms and Finnmark. All pelagic cod were taken in Analysis of variance (ANOVA) was used to test for pelagic trawls at 0 to 60 m depth, in hauls lasting one half effects of group identity, year class, and year of sampling to one hour. on vertebrae number for cod sampled on the spawning BARENTS SEA 500 Tromsøflaket -200 - Ullsfiôrd Tromsø / •70‘ Porsanger Balsfjord Altafjori Varanger Vågsfjord Sørfjori 69' lalanqen 2 00 ' 40 80 km 68' 20 ' 24' Figure 1. Overview of the area sampled 512 S. Løken, T. Pedersen, and E. Berg ICES mar. Sei. Symp., 198 (1994) Lofoten 1991 55 113 54 7 0 93 53 3« 3 0 120 52 Lofoten 1992 Malangen 1992 55 54 54 53 l|4 3» 20 20 250 53 117- 4- 52 50 VERTEBRAE VERTEBRAE NUMBER 52 2 Lofoten 1993 Malangen 1993 55 54 54 53 - - '-2 B 20 -------2 0 03,114 300,00 70 112 20 1 H116 5* 30 52 50 30 52 AGE (years) AGE (years) Figure 2. Comparison of number of vertebrae (average vertebrae number AV ± 2 X SE(AV)) in different year classes of coastal cod (CC) (circles) and Northeast Arctic cod (NA) (squares) sampled on the spawning grounds in Lofoten 1991-1993 and Malangen 1992-1993. Sample sizes are given. ICES mar. Sei. Symp., 198 (1994) Vertebrae numbers as an indicator 513 grounds. Because of the unbalanced data set it was Table 2. Analysis of variance testing the effects of year class necessary to restrict the ANOVA to year classes that (1986-1987), sampling year (1992 and 1993) and sampling were present in all sampling years. A probability level of locality on vertebrae number in coastal cod (CC) sampled at the spawning grounds in Lofoten and Malangen. SS = sum of 5% was taken as indicating statistical significance. All squares, DF = degrees of freedom, MS = mean square, F = statistical tests were performed using SYSTAT (Wilkin­ Fisher ratio, p = significance level, n = 103. son, 1990). Source SS DF MS F p Year class 0.002 1 0.002 0.003 0.959 Results Sampling year 5.938 I 5.938 7.419 0.008 Sampling locality 20.512 1 20.512 25.629 <0.001 Cod sampled on the spawning grounds Error 79.234 99 0.701 In cod sampled on the spawning grounds in Lofoten and Malangen, there was no statistical significant difference in vertebrae number between female and male fish for either NA or CC (ANOVA). ANOVA on the material from Lofoten was restricted The cod juveniles sampled along the coast of Finn­ to the year classes 1983, 1984, and 1985. The ANOVA mark in 1990 at depths of 15-30 m had AV values revealed that there was a highly significant effect of group ranging from 52.6 to 53.0. Samples of settled cod juven­ identity (p < 0.001), a just significant effect of year class iles that were taken at intermediate water depths (26- (p = 0.044), but no significant effect of sampling year in 110 m) in autumn 1991 had AV values ranging from 52.9 the material from Lofoten (Table 1).