The family Gadidae is a group of benthopelagic fishes that inhabit the coastal zones, continental shelves, and slopes of the northern oceans. Gadids comprise more than 50 species in 21 genera, arranged as three subfamilies. Phycines (commonly known as hakes) and lotines (including the freshwater burbot), shown here in red and green, resemble the ancestral form, with one or two dorsal fins, one caudal fin, and elongate, laterally compressed bodies. Gadines, shown here in blue, are the most morphologically derived members of the family, and show greater differentiation of the dorsal and caudal fins.

Gadines are also the more familiar forms, and include such species as codfish, pollocks, haddock, and whiting. These species make the subfamily one of the most important groups of commercially exploited fish, and together they represent more than one-tenth of the annual catch worldwide. For example, the Grand Banks cod fishery, as illustrated here, has sustained the people of  Newfoundland and Labrador for almost 500 years. Despite their commercial importance, the evolutionary relationships of gadids are poorly understood. Most of the species were known to Linnaeus, and many were originally placed in the genus Gadus, but there has since been little agreement on their classification. Most previous work has relied on comparisons of morphology or life history, and has emphasized relationships among families and subfamilies.

One of the outstanding questions is the origin of the present biogeographic distribution of the family. Gadines, the most widely distributed subfamily, have their greatest diversity in the Atlantic Ocean, but several species are endemic to Pacific waters. Today, I want to discuss what we’ve learned about the molecular systematic relationships in the subfamily Gadinae, and in particular, how this explains the origins of the Pacific species.

Collecting localities for 14 species are shown here. The data I will discuss were collected by Dorothy Crutcher and Dave Kivlichan as part of their undergraduate honors theses at the Memorial University of Newfoundland in St. John’s. Dr. Pierre Pepin and I are also pleased to acknowledge the assistance of numerous  international colleagues in collecting these fish, as well as help of the groundfish staff at the Northwest Atlantic Fisheries Centre in St. John’s.

We sequenced two portions of the mitochondrial genome, a 401bp segment at the 5' end of the cytochrome b gene, and a 495bp segment of the cytochrome oxidase I gene, for a total of almost 900 base pairs. DNA was amplified by PCR, and sequencing was done with a fluorescent dye-terminator chemistry and an Applied Biosystems Automated DNA sequencer.

Maximum likelihood, neighbor-joining, and maximum parsimony analyses were performed with version 4 of Dave Swofford’s PAUP program. All methods and variants produced trees with essentially similar topologies. The groups shown here were identified as clusters or clades in a majority of bootstrap replicates from all three methods:

(1) First, All three species of the eastern Atlantic pouts and poutings, genus Trisopterus. This genus is the sister group to the remaining taxa.
(2) Second, The two tomcod species of Microgadus, plus Eleginus navaga. However, the arctic Navaga is more closely related to the Pacific species of tomcod than either is to the Atlantic species.
(3) Third, haddock plus whiting, Melanogrammus and Merlangius respectively.
(4) Fourth, The three species of codfish in the genus Gadus, plus Arctic cod Boreogadus and Walleye or Alaskan pollack Theragra.

These results have a number of interesting implications for the biology of codfish and their relatives, and especially for the relationships and biogeographic origins of the endemic Pacific species.

The genus Gadus presently comprises three nominal species: morhua, macrocephalus, and ogac, the Atlantic, Pacific, and Greenland cod, respectively. Stewart Grant and Gunnar Ståhl have previously provided evidence that Pacific cod migrated through the Bering Strait when it first opened, approximately three to three-and-a-half million years ago. They showed that this species is less genetically variable than Atlantic cod, which is consistent with the notion that it underwent a bottleneck at the time of its origin. As shown here, Greenland cod are usually regarded as more closely related to the partially sympatric Atlantic species than either is to the allopatric, Pacific species.

The first genetic surprise, then, is that Pacific and Greenland cod have identical DNA sequences, for both gene segments. This result has been confirmed with several individuals from both nominal species, and was previously noted in protein electrophoretic comparisons by Dr. Claude Renaud. Analysis of Pacific Cod supplied by Dr. Kenji Saitoh from Japan and Kamchatka shows them to be genetically identical with fish from the eastern Pacific. Pacific and Greenland cod seem to be much less genetically differentiated than transatlantic populations of Atlantic cod, or even populations on the Grand Banks of Newfoundland and the adjacent offshore seamount, Flemish Cap. We therefore suggest that Greenland cod are simply a northward and eastward extension of the range of Pacific cod. As shown here, Greenland cod should be synonymized with Pacific Cod as Gadus macrocephalus. A lower level of cytoplasmic genetic diversity throughout the range of a geographically widespread macrocephalus may stem from the same bottleneck that reduced nuclear gene variation.

Relationships among Gadus, polar cod, and walleye pollock are not yet resolved. However pairwise sequence differences suggest separations at the same time as those within Gadus. Pending clarification of their relative branching order, we suggest that the second endemic Pacific species, walleye pollock (Theragra), is an independent invasion of the Pacific basin. Theragra, like Gadus macrocephalus, is genetically depauperate as compared with Gadus morhua in respect of its protein electrophoretic alleles, and, also like morhua, it is not divided into multiple local stocks. Life history characteristics that have made Atlantic cod and walleye pollock among the world’s most successful fishery resources may well be the result of a common evolutionary history.

The last genetic surpise, for today at least, concerns the origin of Pacific tomcod. The molecular data indicate that the Pacific tomcod is more closely related to Eleginus than it is to the other, congeneric species, the Atlantic tomcod. The genetic distance between Eleginus and Microgadus proximus is approximately equal to that between Pacific and Atlantic cod. Then, a molecular clock would indicate that this third endemic Pacific species was derived from Eleginus at about the same time as the separation of the two species of Gadus. The ancestor of  Microgadus proximus evidently evolved in the polar basin and dispersed through the Bering Strait into the Pacific Ocean. To keep Microgadus monophyletic, Eleginus navaga should be included in that genus.

The relationships of the saffron cod, Eleginus gracilis, which are currently under investigation, help clarify this picture. This species occurs mainly in the western Pacific, but overlaps extensively with Microgadus proximus in the waters around the Aleutian Islands south of Alaska. The most likely hypothesis is that this distribution represents a fourth independent invasion of the Pacific Basin. Preliminary molecular data indicate that the two species of Eleginus are in fact sister species. The restricted distribution of navaga in the vicinity of the White Sea could then be seen as a relatively recent, westward expansion from the more widespread gracilis. Such a relationship is also favored by morphological analysis. The two Eleginus species share a distinctive skeletal feature, an expansion of the tips of the precaudal vertebrae. The condition of gracilis is intermediate between Microgadus and navaga, which also suggests that navaga is derived from gracilis. Finally, Microgadus and Eleginus are distinct from all other gadines in possessing continuous lateral lines along all or part of their length. This kind of reasoning shows that molecular and morphological analyses can be complementary, not antagonistic.

This molecular analysis will help to clarify the morphological evolution of gadines. In the classical study of gadiform evolution, Svetovidov in 1948 constructed a key to gadid genera that can be interpreted as a phylogenetic hypothesis. As shown here, each successive taxon, moving from the left to right, is the sister to the remaining taxa. In common with our study, Svetovidov recognized Trisopterus as the outgroup to the remaining gadines. He also recognized the similarity of Microgadus and Eleginus by arranging them as successive taxa in his key, although they are therefore not each other’s closest relatives. However, Svetovidov did separate Gadus from Boreogadus and Theragra, which is inconsistent with the molecular data.

The most recent analysis of gadid relationships is that of Dr. Jean Dunn, who performed a cladistic analysis of 28 morphological and osteological characters among eleven species, including nine in the present study. Except for close relationships among Gadus, Theragra, and Boreogadus, there is little or no similarity with the evolutionary relationships implied by the molecular data. For example, there are no indications of a close relationship between Eleginus and Microgadus. For the taxa common to both studies, Dunn’s morphological hypothesis would require 30 changes, whereas  the molecular trees presented here would require at most only an extra 5 events. In contrast, molecular trees equivalent to those presented here would require no more than 450 mutations, whereas the morphology tree equivalent to Dunn’s hypothesis would require at least 52 additional mutations. We conclude that, although morphological data are compatible with the molecular tree, the reverse is not true.

In summary, molecular analysis suggests that codfish and their relatives in the subfamily Gadinae have their biogeographic origins in the eastern Atlantic, as represented here by the genus Trisopterus. Three species of gadids endemic to Pacific waters, shown here in red, that is, the Pacific tomcod, Pacific cod, and Walleye pollock, represent parallel radiations into the Pacific, all of which occurred more or less simultaneously with the first opening of the Bering Strait, three to three-and-a-half million years ago. Greenland cod appear to be a recent, northward and eastward extension of Pacific cod, Gadus macrocephalus, and should be synonymized with that species. The fourth Pacific species, saffron cod, and its arctic sister species navaga in the genus Eleginus are part of a common evolutionary lineage with Pacific tomcod, which should be synonymized under the name Microgadus.

At Memorial University, we are currently working to expand this study to include other genera of gadiform fishes. For example, work in Heather Warren’s honours thesis indicates that Lotines, rather than Phycines, are the immediate sister group to the Gadines, and that Merluccid hakes are the sister family to the Gadidae.