Download as pdf or txt
Download as pdf or txt
You are on page 1of 10

Reviews in Fish Biology and Fisheries 11: 171–179, 2002.

© 2002 Kluwer Academic Publishers. Printed in the Netherlands.


171

The origin of the salmonid fishes: marine, freshwater . . . or neither?

R.M. McDowall
National Institute of Water and Atmospheric Research, PO Box 8602, Christchurch, New Zealand (E-mail:
R.McDowall@NIWA.CRI.NZ)

Accepted 10 April 2002

Contents
Abstract page 171
Introduction 171
Diadromy in Salmonidae
An historical perspective on opinions regarding the origins of the Salmonidae 172
The geochronological place of diadromous species in salmonid history 174
Diadromy and phylogenetic analysis 174
Origin of Salmonidae in a cladistic context 175
Relationships of salmonids in the context of variability of diadromous life cycles 176
Conclusion 177
Acknowledgements 178
References 178

Key words: diadromy, freshwater, marine, origins, Salmonidae

Abstract

The evolutionary origins of the salmonid fishes, whether in freshwater or the sea, have been debated for
centuries. Early views favoured a group of marine ancestry invading freshwaters; more recently, there was a shift
towards a freshwater ancestry, on grounds that a return to freshwater to spawn indicates the ancestral biome.
Salmonids are widely believed to share an ancient common ancestry with the northern hemisphere Osmeridae
and southern hemisphere Retropinnidae and Galaxiidae. Salmonidae are diadromous, as are Osmeridae,
Retropinnidae and Galaxiidae. This suggests that diadromy is an ancient behavioral phenomenon across all
these groups, that the shared common ancestry of these groups was also diadromous, and that the ancestry of
Salmonidae was neither marine nor freshwater, but was amongst diadromous fishes. This begs the question of
whether this common ancestor was marine or freshwater, a question for which an answer seems likely to be elusive.

Introduction ancestry during more recent decades (see below).


The question has been so intriguing probably because
The question of whether fishes of the family of the apparently deep-seated (and usually correct)
Salmonidae have a marine or freshwater ancestry is notion that freshwater fishes are generally unable to
a venerable one that has been discussed often and make a transition to marine habitats (or marine fishes
for decades. It has attracted the attention of some make the reverse transition – though this seems to have
great ichthyologists of the past, and continues to been of less interest) and yet salmon and trout can
attract comment. Opinions, summarized across nearly move, seemingly freely, between these two biomes.
two centuries, seem evenly divided over the answer, That there has been so much debate about salmonid
though perhaps support has shifted from a majority origins is perhaps enigmatic, given lesser interest in
early belief in a marine origin, towards a freshwater the same question for other fish that move equally
172

freely between marine and freshwater habitats – such phases (varying among species) in freshwater
as lampreys (Petromyzontidae, Geotriidae), sturgeons before they emigrate to sea. Most feeding and
(Acipenseridae), anguillid eels (Anguillidae), smelts growth take place at sea, usually over several
(Osmeridae, Retropinnidae), shads (Clupeidae), and years. Approaching maturity, the adults return to
so on (see McDowall, 1988). freshwater, and penetrate up rivers to breeding
Tchernavin (1939, p. 120) addressed the question grounds, where they spawn.
for salmonids, suggesting that the many questions A common feature of anadromous fishes is
arising from consideration of the salmon life history homing (a return to their natal habitat), to spawn,
“finally resolve into the cardinal one: the origin of a habit that may have synergies with anadromy, in
salmon. One of its two present haunts must have enabling them to re-locate and exploit historically
been the original habitat of the salmon’s ancestors” productive spawning habitats (McDowall, 2001).
(original italics). Thorpe (1982) concurred, consid- A review of diadromous fishes (McDowall, 1988)
ering the “central question” to be: “did the salmonids showed that there are about 227 species, of which
originate in the sea or in fresh water?” Stearley 110 are anadromous, with 27 of these being salmonids
(1992, p. 662) thought this a “long-standing contro- (including Coregonus, which is sometimes placed in a
versy”, and found no “unanimity of opinion on the separate family). Anadromy is thus a well-represented
ancestral condition and subsequent historical trends in behavior amongst salmonids and they constitute a
the life histories for the Salmonidae.” Balon (1968, significant proportion of all known anadromous fishes.
p. 6), addressing the question at a much narrower There are many non-anadromous salmonids, and
focus, asked: “What is the original form – brown similarly, many non-anadromous stocks of otherwise
or sea trout?” and concluded that “anadromous trout anadromous species. Given these interrelated facts,
have evolved gradually from an unknown fresh-water and the additional fact that salmonids are amongst the
ancestor.” But, in an apparent contradiction (p. 9), he most important and well-studied of all fish, owing to
also thought that “the recent form of salmon and trout their high profile in both commercial and recreational
are obviously descendants of the anadromous stock, fisheries, perhaps it is no surprise that the question of
irrespective of their possibly marine or fresh-water their origins has commanded so much attention.
distribution.”
In this discussion paper, I revisit the question
of salmonid origins, adding an additional, perhaps An historical perspective on opinions regarding
peculiar, rider that salmonid origins may be neither the origins of the Salmonidae
freshwater nor marine.
Opinion about salmonid origins, whether marine or
Diadromy in Salmonidae freshwater, date back at least to the early 19th century,
leading Tchernavin (1939, p. 121) to despair after
Diadromy is a general term introduced to the ichthy- reading the available literature. A very early English
ological lexicon by American ichthyologist George commentator (Pennant, 1776), is cited by Day (1887)
Myers (1949). Diadromous fish are those that migrate as apparently having a neutral view, simply stating that
on a regularly-timed, consistent, and most often oblig- “the salmon is a fish that lives both in the salt and fresh
atory basis between freshwaters and the sea. Myers waters, quitting the sea at certain seasons for the sake
defined several sub-categories of diadromy, including of depositing its spawn in security.” Fleming (1828,
the already long used terms anadromy and catadromy, p. 179) is said by Day to have considered the salmon
and he coined the new term amphidromy. Each a “migratory fish from the sea”, perhaps implying a
term covers different sub-categories of diadromy (see marine origin, though this is unclear. And Day (1887)
McDowall, 1999 for refined definitions). In general, reported Parnell (1838, p. 279) as arguing that “there
any salmonids that are diadromous fall into the sub- is no doubt that the true abode of the salmon is in the
category anadromous (but see Stearley, 1992). sea, for as soon as it has entered the rivers it begins
Anadromy has the following general features: to deteriorate in condition.” Day (1887) and Frank
Spawning is in freshwater, and the newly hatched Buckland, according to Day, held contrary views. For
larvae remain there for periods varying from a example, Day (1887) cites Buckland (1873) as consid-
few days to a few months (occasionally a few ering Atlantic salmon (Salmo salar) “a sea fish proper;
years). They feed and grow through some juvenile nevertheless, this fish ascends the rivers and streams
173

in order to deposit its eggs, for unlike other sea fish, 5. Existence of a homing migration to fresh water
it does not breed in the sea.” But their compatriot, indicates the origin [source] of the migratory
British Museum ichthyologist Albert Günther (1866, habit.
p. 3) described salmon as a freshwater fish that has 6. Young Salmonidae show “amazing” resemblance
adopted the habit of descending to the sea to feed. to the adults of freshwater forms.
C.T. Regan (1911, p. 24), also at the British Museum Various subsequent commentators have followed
(during the early 1900s), and one of the more notable Tchernavin’s view, for various reasons (e.g., Vlad-
of early 20th century ichthyologists, thought that ykov, 1963). Hoar (1976, p. 1235) carefully re-
salmonids “may be regarded as marine fishes [that] evaluated Tchernavin’s arguments and concluded that
are establishing themselves in fresh water.” Meek “they still make sense and nothing seems to have been
(1916, p. 115) agreed, presenting arguments “why discovered in the past half century that invalidates
the evolution [of salmonids] is assumed to be from the hypothesis of a freshwater origin.” Hoar (1976,
marine to fresh-water conditions. . . . The reason . . . p. 1236) further argued that additional information on
is that the Atlantic salmon and the trouts are more distribution, morphology, biochemistry, physiology,
closely related than the Atlantic and Pacific salmons, and behavior “seem to have added support to most of
that in other words the salmons preceded the trouts.” [Tchernavin’s] arguments.”
Boulenger (1917), who was at the British Museum While some found Tchernavin’s arguments
with Regan, also supported a marine origin. Thus, persuasive, Thorpe (1982, p. 88) described them as
opposing views amongst contemporary biologists are “argued forcibly and authoritatively” but (Thorpe,
evident from early times, although there appears to 1988) “with more conviction than evidence” – I
have been a minor gathering of “authority” around think with some validity. Not one of the points listed
a marine ancestry in the late 19th and early 20th above from Tchernavin (1939) seems, to me, to
centuries. indicate a freshwater ancestry. Salmonids are, of
Similarly divergent opinions continued into course, well-adapted to spawning and early growth
modern ichthyological history. Unlike many earlier in fresh water – where they have been spawning and
authorities discussed above, Tchernavin (1939) undergoing early life for millions of years. And why
decided on a freshwater origin. His analysis of the should what salmonids do now indicate what their
marine/freshwater ancestry of salmon (Salmo salar) ancient ancestors did before them? Thus, Tchernavin’s
is certainly the most cited and carefully developed (1939) argument rests substantially on false inference
perspective from the mid 20th century until recent from contemporary importance to historical priority
times. That being so, it is worth summarising (Gould, 1983).
Tchernavin’s argument: It is also unclear why a migratory life history
should be derived from a non-migratory one, rather
1. Many genera are wholly freshwater but none is
than the reverse. This is especially true when it
entirely marine, and the great likeness between
is recalled that loss of diadromous migrations has
freshwater and migratory forms of salmonids
recurred repeatedly in diverse species of extant
suggests recent evolution of migratory forms.
Salmonidae (and other diadromous families). If this
2. Those salmonids that are the least marine are is indicative of what happened in salmonid history,
the most primitive, the trout (which Tchernavin Tchernavin’s (1939) assumption of migratory species
wrongly regarded as non-migratory) being more being recently derived is difficult to accept.
primitive than the salmon (migratory). Thorpe (1982) presented an explicit argument in
favour of a marine ancestry, and as such his work
3. The distribution of salmonids resembles that of
also needs discussion. Thorpe (1982, p. 89) observed
freshwater plants and animals.
that there is a shift from diadromy (anadromy) to
4. The migratory salmon expends great effort in landlocked/freshwater populations in some salmonid
getting into freshwater to spawn, finding condi- species, and argued that therefore:
tions congenial for the activity of eggs and sperm,
for fertilisation, and for development and early the evolutionary trend in the family appears to
life of the young; moreover persistence, genera- be toward total dependence on fresh water, and
tion after generation, of salmon returning to fresh away from dependence on the sea. It is then appro-
water indicates their origins. priate to view the Salmonidae as relatively prim-
174

itive teleosts of probable marine pelagic origin” stocks of salmonids “during the Glacial period”,
. . . meaning that “. . . the fresh water phases of so that “divergence into freshwater and migratory
the life histories of salmonids are enabled by forms occurred only recently.” Clemens (1953) and
specific adaptations, which are superimposed on Neave (1958) similarly suggested that Oncorhynchus
a basically marine organism. diverged from Salmo as recently as Pleistocene glacial
Thorpe thus viewed salmonids as evolving away times. The logic for these assertions is, at best,
from a diadromous (anadromous) (marine/freshwater) elusive.
life to a purely freshwater one – reflecting Regan’s Apart from the fact that the amount of diversific-
(1911) view that salmonids may be regarded as marine ation and geographical spread represented by modern
fish establishing themselves in freshwater. Thorpe’s diadromous salmonids seems intuitively inconsistent
argument seems, to me, to also be erroneous on at with such a recent origin, these notions are clearly fals-
least two grounds. Firstly, though not necessarily ified by the fossil record. Cavender and Miller (1972)
most importantly, Thorpe’s (1982) argument (as with described Smilodonichthys rastrosus, a huge Pliocene
Tchernavin’s) rests on false inference from contem- salmonid fossil from western North America, (now
porary importance to historical priority (Gould, 1983). included in Oncorhynchus – Stearley and Smith, 1993,
Secondly, he appears to be arguing that future evolu- p. 23), and thought this fish diadromous (anadromous)
tionary directions of salmonids were determined by based on the site from which the fossil came, the
the direction of their evolutionary past, thus implying likelihood that it was a pelagic filter feeder (it had
that the evolution of salmonids has momentum and numerous gill rakers), its large premaxillary teeth, and
direction, projecting adaptation forwards. This is an internal organisation of the dermal bone (cavernulous
essentially orthogenetic argument, a form of evolution providing potential storage of fats – p. 37).
that is generally rejected by conventional evolutionary Cavender and Miller (1972, p. 39) also reported
biologists (Mayr, 1982). “Fragmentary remains of what appears to be an
Thorpe (1982) and McCormick and Saunders extinct species of Oncorhynchus” from western North
(1987) listed Balon (1968) as one contrary author America (of Pliocene age), leading them to conclude
favouring a marine ancestry. Thorpe (1982: 93) that the genus is “at least as old as middle Plio-
summarized Balon’s (1980) hypothesis of salmonid cene.” Moreover, Stearley (1993, p. 23) cited several
evolution as “evolution of life history strategies other Tertiary fossils, the most telling of which (with
through penetration of fresh water by a pelagic marine regard to a supposed Pleistocene origin of diadromous
fish, and progressive restriction of life-history to the salmonids), being a Miocene fossil of chum salmon,
freshwater habitat”, but this does not seem to be Oncorhynchus keta, from western North America.
Balon’s position. Balon (1980, p. 552) wrote that Stearley and Smith (1993, p. 23) found that “these
“evidence of the evolution of reproductive guilds . . . fossils of Oncorhynchus document abundant evidence
seems to support [the] assumption . . . that earlier for a minimum age of 6 million years for the modern
forms [of salmonids] were related to the oceanic pela- species of Pacific salmons and trouts”, and accept-
gial by their reproductive strategy”, which Thorpe ance of this age certainly lays to rest any notion
(1982) took to mean “evolution of salmonid life that a Pleistocene diversification produced these taxa.
history strategies through penetration of fresh water There seems good evidence for probably anadromous
by a pelagic marine fish.” It seems to me that Thorpe salmonids as early as the Miocene, far earlier than
misinterpreted Balon’s suggestion that freshwater speculated by Tchernavin (1939) and those who have
populations of salmonids evolved from anadromous followed him.
populations, as Balon’s support for a marine ancestry.

Diadromy and phylogenetic analysis


The geochronological place of diadromous species
in salmonid history Diadromy is a behavioral trait, and there is a view that,
as behavioral traits tend to be both non-quantitative
Two distinctly contrasting views have been expressed and malleable, they are of dubious value in phylogen-
regarding when the first diadromous salmonids etic analysis. However a converse view has prevailing
evolved. Tchernavin (1939), believing in a fresh- acceptance, viz. that when applied with care, behavi-
water origin, argued for development of sea-migratory oral traits can be used in phylogenetic analysis with
175

useful results (Dobson, 1985; McLennan et al., 1988; populations have retained the instinct to migrate to
de Quieroz and Wimberger, 1993; Paterson et al., sea (Northcote, 1992; Jonsson and Jonsson, 1993).
1995; McDowall, 1997). Nordeng (1983) has shown this for Arctic char
Diadromy poses problems for phylogenetic (Salvelinus alpinus) and other instances include brown
analysis, particularly though not exclusively amongst trout when non-migratory stocks were translocated in
salmonids. For example, many diadromous salmonids Europe (Thorpe, 1987), and when introduced to the
also have non-diadromous populations – landlocked Falkland Islands (Nolan, 1993).
populations establish and forsake migrations to With all this within-species variability, it appears
and from the sea, sometimes replacing these that diadromy is a trait of limited utility in phylo-
with migrations to and from lakes. Perhaps best genetic analysis. Despite this, one notable aspect of
known are kokanee populations of sockeye salmon diadromy is that it has evolved relatively rarely across
(Oncorhynchus nerka) in western North America the diversity of fishes. Diadromy is known in only 36
(Foerster, 1968; Foote et al., 1994; Wood and of 482 families of fishes (McDowall, 1988; Nelson,
Foote, 1996), but there are numerous other examples 1994). Most diadromous fish species are found in only
amongst salmonids (and elsewhere amongst the a few families – in the listing of McDowall (1988),
related Osmeroidei – McDowall, 1988). Rounsefell there are 11 sturgeons, 27 salmonids, 11 salangids,
(1958), for instance, actually ranked salmonids of 30 clupeids, 15 anguillid eels, 15 mugilid mullets,
western North America according to the degree to 64 gobies. Thus 173 of the 227 species, or 76% of
which their diadromy is facultative. species listed in McDowall (1988) are in only seven
It is known that in some normally diadromous families. Diadromy may have evolved several times
species, a small proportion of individual fish may in some of these families but, even so, it is clear
be non-diadromous (they fail to migrate to sea), that diadromy is a trait that has evolved relatively
though they may participate in reproduction with rarely. This rarity, together with variability in the pres-
their diadromous con-specifics at spawning time. This ence/absence of diadromy within species discussed
may occur in both sexes, but is reported especially above, suggests that, at best, diadromy may be of
for males of species of Salmo, Oncorhynchus and phylogenetic interest when comparing higher taxa that
Salvelinus (Balon, 1968; Saunders et al., 1982; Silver- do, or do not, exhibit diadromy. The evolution of
stein and Hershberger, 1992; Jonsson and Jonsson, diadromy may be a novel and phylogenetically signifi-
1993; Unwin et al., 1999; Healey et al., 2000). cant event, and its presence in two or more species in
The presence of these few non-migratory indi- a family that is otherwise non-diadromous may hint at
viduals among the many became evident when relationships of those where it is present. But, owing
several diadromous salmonids were introduced to to within and between taxon variability, differences in
New Zealand in the late 1800s and early 1900s. the nature of the diadromy (or its presence/absence
Diadromous stocks of sockeye salmon, Atlantic in species being compared) seem of minimal phylo-
salmon, and rainbow trout (O. mykiss) were trans- genetic value when exploring the relationships of
located to New Zealand, but in all three instances congeneric or confamilial species.
the outcome was acclimatization of entirely non-
migratory stocks. There must have been some indi-
viduals amongst the translocated fish that had a genetic Origin of Salmonidae in a cladistic context
tendency to not migrate, and these formed the ancestry
of populations that became established (McDowall, Few of the authors discussed earlier have examined
1990, 1994). Instances are known of latitudinal vari- the question of salmonid relationships by identifying
ation in the extent to which populations of a species common derived characters, i.e., from a cladistic
are diadromous, typically with a greater tendency for perspective. Stearley and Smith (1993, p. 5) recog-
diadromy at higher (cooler latitudes), as with intro- nised that Regan (1914) used “shared specializa-
duced brown trout (Salmo trutta) in New Zealand tions as evidence of relationship as later formalized
(McDowall, 1990). by Hennig (1966)”, whereas (p. 4) they found that
Conversely, there are records of non-diadromous Tchernavin (1939) used “shared primitive characters
populations of diadromous species being translocated for diagnosing genera”, which provide no indica-
to new locations and reverting to diadromy, indic- tions of relationships. Thus, virtually all early studies
ating that some individuals in such non-migratory searched for characters that indicate a marine or fresh-
176

water ancestry using “indirect” methods (Stearley, specified only Thaleichthys), since these are regarded,
1992, p. 622), the characters having no obvious indic- by some, as the most basal (plesiomorphic) of
ation of why a marine or freshwater origin should osmerids. He addressed the “span of morphology
be indicated – there is no reason why contemporary defining ancestral states for salmonid characters”, and
character states should indicate which of these is used the phylogeny generated for interpreting the
true, as Tchernavin (1939, p. 127) admitted when history/evolution of diadromous migratory strategies
noting that “the distribution of an animal does not in the subfamily. He argued that in Salmoninae, in
always follow the milieu inhabited by its ancestors” particular, a series of non sea-migratory genera such
(though he appears to have taken no notice of his as Brachymystax, Acantholingua, Salmothymus and
own warning). For example, it is widely suggested Platysalmo are more plesiomorphic sister-groups of
that because salmonid reproductive products survive diadromous (anadromous) apomorphic genera such as
and result in fertilisation only in freshwater, and Salmo, Salvelinus and Oncorhynchus. On this basis,
that because egg development takes place there, a Stearley (1992) argued for a freshwater origin of
freshwater origin is indicated (Tchernavin, 1939; salmonines because the genera listed above as “prim-
McDowall, 1988), and this argument looks good, itive” are non-migratory. Greenwood (1993, p. 374)
as long as you don’t look too carefully. Of course described Stearley’s (1992) account as a “new and
salmonid eggs and sperm function best in freshwater convincing solution to the old problem of marine
if this is where salmonid reproduction has taken versus freshwater ancestry for salmonoids”, though
place for the past several (many?) millions of years, in fact Stearley (1992, p. 623) wrote explicitly of
and this scarcely argues for a freshwater origin at his “focus on the life histories of the salmonine
some long-past time. The present character state does subfamily”, which is rather different (emphasis added
not prevent some change of reproductive medium in in both quotes).
the past, the possibility of which is explicitly indi-
cated by the fact that although most species in the
related family Osmeridae spawn and develop in fresh- Relationships of salmonids in the context of
water (like Salmonidae), three species of osmerid, variability of diadromous life cycles
in three different genera spawn and develop at sea.
Two of these genera (Hypomesus and Spirinchus) have Although Stearley (1992) anchors his morpholo-
both marine and freshwater spawning species (McAl- gical phylogenetic analysis in outgroups, rather than
lister, 1963). Given the varied spawning media in carrying this approach through into comparable iden-
Osmeridae, assumptions that Salmonidae originated tification of outgroup character states in his behavioral
in freshwater because they spawn and develop there analysis, he uses the phylogenetic relationships to
are fraught with risk. Unlike these marine osmerids, determine patterns of behavioral evolution amongst
no salmonids have adopted entirely marine life history the taxa. This forms the basis for his conclusion
strategies, pink salmon (Oncorhynchus gorbuscha) that salmonines have a freshwater ancestry. Stearley’s
being closest with their spawning, sometimes within (1992) argument for a freshwater origin of salmonines
tidal influence in low elevation rivers (Heard, 1991). is suspect for several reasons:
Stearley (1992) addressed the origin of the 1. As discussed above, diadromy is a notori-
subfamily Salmoninae in an explicitly cladistic ously variable or reversible character trait within
analysis of the constituent genera and species – he species, and that being so, specification of non-
considered (p. 622) that a “good understanding of diadromy in salmonines as plesiomorphic and
salmonid phylogenetic relationships is of extreme diadromy apomorphic is unjustified.
import for understanding the evolution of their life 2. Stearley (1992, p. 627) argues that “Primitive
histories”, hoping that, structured upon a proper salmonid sistergroups – Prosopium, Thymallus
framework, “the sequence of branching relationships and the archaic trouts [those listed above
of taxa . . . a sequential elaboration of salmonid beha- as the primitive salmonines] are restricted to
viors may be discernible.” I agree. freshwaters”, whereas “Advanced salmonine
Stearley (1992, p. 625) stated that he constructed and coregonine clades – Coregonus, Stenodus,
his phylogenetic relationship using (amongst other Salvelinus, Hucho, Salmo and Oncorhynchus
taxa) the osmerid genera Thaleichthys and Spirin- all contain sea-run species”. He regarded non-
chus as outgroups (Stearley and Smith, 1993, p. 7, migratory Salvelinus and Hucho species as
177

“atavistic, reverting to the primitive (strictly and galaxioid families, including most species
freshwater) condition possessed by more prim- of Osmeridae (including Plecoglossus), Salangidae
itive salmonine sister groups.” But there is (both Osmeroidei) and Retropinnidae (including
no reason why the primitive salmonine sister Prototroctes) and Galaxiidae (including Aplochiton
groups (Brachymystax, Acantholingua, Salmo- and Lovettia) (both Galaxioidei), i.e., diadromy
thymus and Platysalmo) should not also have is widely present in a recognised sister-group
lost their formerly diadromous behaviors just of Salmonidae, suggesting that that diadromy
like these Salvelinus and Hucho species have (as was a characteristic of the group/species ancestral
also have various species of Salmo and Onco- to the entire [Salmonidae+Osmeroidei+Galaxioidei]
rhynchus); Stearley (1992, p. 629) himself notes complex. That being so, there are no grounds
that “Spontaneous abandonment of migratory life for either regarding non-diadromous salmonids as
cycles appears to occur regularly” in salmonines. plesiomorphic (indicating the ancestral condition), or
So, it is unclear why should this not have for regarding the origins of Salmonidae as being in
occurred in the listed primitive salmonines whose fresh water (recall that although most of these taxa
lack of diadromy is interpreted by Stearley as are freshwater spawners, three osmerids are marine
plesiomorphic and indicating a freshwater origin spawners).
in salmonines.
3. In spite of the above proposed set of relation-
ships, looking more widely at the Salmonidae Conclusion
(i.e., including coregonines and thymallines),
according to Stearley and Smith (1993) the most Prolonged debate has centered around whether the
plesiomorphic genera in the family are, Core- Salmonidae had a marine or freshwater ancestry, with
gonus, Stenodus and Prosopium; species of Core- strong arguments presented in favour of each. If it is
gonus and Stenodus are diadromous (McDowall, accepted that Osmeroidei+Galaxioidei are the sister-
1988), so that diadromy is present in these group of Salmonidae, however, given the widespread
plesiomorphic genera, creating a conflict with presence of diadromy in Osmeroidei+Galaxioidei,
Stearley’s (1992) conclusion that absence of then it has to be concluded that salmonids had a neither
diadromy in the most primitive salmonines indi- marine nor freshwater origin, but are part of a diverse
cates a freshwater ancestry for salmonines. group of fishes in which diadromy is an ancestral char-
4. Perhaps more importantly than any of the above, acter state (McDowall, 1993). This conclusion, of
relationships of the Salmonidae, and the signal course, begs the question of whether the broader group
indicated about the family’s origins (marine comprising Salmonidae+Osmeroidei+Galaxioidei had
or freshwater), must be viewed in an even a marine or freshwater origin. I know of no way this
broader context. Specifically, although details are broader question can be answered with any assurance
debated, there is quite wide and long-standing or finality. Flexibility and reversibility in the pres-
acceptance that Salmonidae are a sister-group ence/absence of diadromy mean that we may never
of Osmeroidei+Galaxioidei (McDowall, 1969; know whether Salmonidae+Osmeroidei+Galaxioidei
Rosen, 1974; Fink and Weitzman, 1982; Johnson had a marine or freshwater ancestry. Fyhn et al. (1999)
and Patterson, 1996 – family assignments and have explored origins of teleost fishes as a whole,
arrangements vary, but I think the taxa involved based on views that the radiation of teleosts began
in these relationships are quite clearly under- in fresh water. They suggested that the “first trials
stood). That this view is accepted, at least in of colonising the saline ocean was by way of marine
part, by Stearley (1992) and Stearley and Smith’s excursions into the sea for feeding, and return to fresh
(1993) is shown by their choice of the osmerids water for spawning.” As they point out, anadromous
Spirinchus and Thaleichthys as being amongst the fish “such as the salmonids still show this behavior.”
outgroups for their salmonid phylogenetics. Perhaps this is a pointer to a freshwater ancestry of the
If it is accepted that [Osmeroidei+Galaxioidei] is entire osmeroid/galaxioid/salmonid complex of fishes.
the sister group of Salmonidae, then the life history On the other hand, it might be argued that fishes of
character states within the [Osmeroidei+Galaxioidei] marine origin invaded freshwater (and there are plenty
ought to be informative of the ancestral condi- of precedents for this across the diversity of fishes),
tion. Diadromy is widely recognised within osmeroid found this a fruitful biome in which to live, and so
178

radiated and expanded to produce the diversity that we Greenwood, P.H. (1993) Review of: Systematics, historical ecology
see today. and North American freshwater fishes. In: Mayden, R.L. (ed.),
Rev. Fish Biol. Fish. 3, 373–375.
Günther, A. (1866) Catalogue of Fishes in the British Museum 6,
373–375.
Healey, M.C., Henderson, M.A. and Burgetz, I. (2000) Precocial
Acknowledgements
maturation of male sockeye salmon in the Fraser River, British
Columbia, and its relationship to growth and year-class strength.
I am grateful to John Thorpe, University of Glasgow, Can. J. Fish. Aquat. Sci. 57, 2248–2257.
Heard, W.R. (1991) Life history of pink salmon (Oncorhynchus
Scotland, Eugene Balon, University of Guelph,
gorbuscha). In: Groot, C. and Margolis, L. (eds.), Pacific Salmon
Guelph, Ontario, Canada, Don Jellyman, NIWA, Life Histories. University of British Columbia, Vancouver,
Christchurch, and also Gerald Smith, University of pp. 119–230.
Michigan, Ann Arbour, Michigan, U.S.A., and a Hennig, W. (1966) Phylogenetic Systematics. University of Illinois,
Urbana, 263 pp.
second anonymous referee, for thoughtful comment
Johnson, G.D. and Patterson, C. (1996) Relationships of lower
on this paper. euteleostean fishes. In: Stiassny, M.L.J. Parenti, L.R. and
Johnson, G.D. (eds.), Interrelationships of Fishes (2nd ed.).
Academic, New York, pp. 251–332.
Jonsson, B. and Jonsson, N. (1993) Partial migration: niche shift
References versus sexual maturation in fishes. Rev. Fish Biol. Fish. 3, 348–
365.
Balon, E.K. (1968) Notes to the origin and evolution of trouts Hoar, W.S. (1976) Smolt transformation: evolution, behaviour and
and salmons with special reference to the Danubian trouts. Vést. physiology. J. Fish. Res. Bd. Can. 33, 1233–1252.
Cesk. Spol. Zool. (Act. Soc. Zool. Bohem.) 32, 1–21. McAllister, D.E. (1963) A revision of the smelt family Osmeridae.
Balon, E.K. (1980) Early ontogeny of the lake charr, Salvelinus Nat. Mus. Can. Bull. 191, 1–53.
(Cristivomer) namaycush. In: Balon, E.K. (ed.), Charrs, McCormick, S.D. and Saunders, R.L. (1987) Preparatory physiolo-
Salmonid Fishes of the Genus Salvelinus. Dr W. Junk, The gical adaptations for marine life of salmonids: osmoregulation,
Hague, pp. 485–562. growth and metabolism. Am. Fish. Soc. Symp. 1, 211–229.
Boulenger, G.A. (1917) Sur l’origine marine du genre Salmo. McDowall, R.M. (1969) Relationships of galaxioid fishes, with a
Compt. Rend. l’Acad. Sci., Paris 164, 1041–1044. further discussion of salmoniform classification. Copeia 1969,
Buckland, F. (1873) Familiar History of British Fishes. Society for 796–824.
Promoting Christian Knowledge, London. 420 pp. McDowall, R.M. (1988) Diadromy in Fishes: Migrations between
Cavender, T.M. and Miller, R.R. (1972) Smilodonichthys rastrosus – Freshwater and Marine Environments. Croom Helm, London,
a new Pliocene salmonid fish. Bull. Mus. Nat. Hist. Univ. Oregon 308 pp.
18, 1–44. McDowall, R.M. (1990) New Zealand Freshwater Fishes: A Natural
Clemens, W.A. (1953) On some fundamental problems in the history and Guide. Heinemann Reed, Auckland, 553 pp.
biology of Pacific salmons. Trans. Roy. Soc. Can. 47 (Ser. III, McDowall, R.M. (1993) A recent marine ancestry for diadromous
Sec. 5), 1–13. fishes? Sometimes yes, but mostly no. Environ. Biol. Fish. 37,
Day, F. (1887) British and Irish Salmonidae. Williams and Norgate, 329–335.
London. 298 pp. McDowall, R.M. (1994) Gamekeepers for the Nation: the Story of
de Queiroz, A. and Wimberger, P.H. (1993) The usefulness of New Zealand’s Acclimatisation Societies, 1861–1990. Canter-
behaviour for phylogeny estimation: Levels of homoplasy in bury University, Christchurch, 508 pp.
behavioural and morphological characters. Evolution 47, 46– McDowall, R.M. (1997) The evolution of diadromy in fishes (revis-
60. ited) and its place in phylogenetic analysis. Rev. Fish Biol. Fish.
Dobson, F.S. (1985) The use of phylogeny in behaviour and ecology. 7, 443–462.
Evolution 39, 1384–1388. McDowall, R.M. (1999) Different kinds of diadromy: different
Fink, W.L. and Weitzman, S.H. (1982) Relationships of stomii- conservation problems. ICES J. Mar. Sci. 56, 410–413.
form fishes (Teleostei), with a description of Diplophos. Bull. McDowall, R.M. (2001) Anadromy and homing: two life history
Mus.Comp. Zool., Harv. Univ. 150, 31–93. traits with adaptive synergies in salmonid fishes. Fish Fish. 2,
Fleming, J. (1828) A History of British Animals. Edinburgh and 78–85.
London. 565 pp. McLennan, D.A., Brooks, D.R. and McPhail, J.D. (1988) The
Foerster, R.E. (1968) The sockeye salmon Oncorhynchus nerka. benefits of communication between comparative ethology and
Bull. Fish. Res. Bd Can. 162, 1–422. phylogenetic systematics: a case study using gasterosteid fishes.
Foote, C.J., Mayer, I., Wood, C.C., Clarke, W.C. and Black- Can. J. Zool. 66, 2177–2190.
burn, J. (1994) On the developmental pathway to nonanadromy Mayr, E. (1982) The Growth of Biological Thought. Belknap,
in sockeye salmon, Oncorhynchus nerka. Can. J. Zool. 72, Cambridge, 974 pp.
397–405. Meek, A. (1916) The Migrations of Fishes. Arnold, London, 437 pp.
Fyhn, H.J., Fin, R.N., Reith, M. and Norberg, B. (1999) Yolk protein Myers, G.S. (1949) Usage of anadromous, catadromous and allied
hydrolysis and oocyte free amino acids as key features in the terms for migratory fishes. Copeia 1949, 89–97.
adaptive evolution of teleost fishes to seawater. Sarsia 84, 451– Neave, F. (1958) The origin and speciation of Oncorhynchus. Trans.
456. Roy. Soc. Can. 52 (ser 3), 25–39.
Gould, S.J. (1983) What, if anything else, is a zebra. Nat. Hist. (New Nelson, J.S. (1994) Fishes of the World (3rd ed.). Wiley, New York,
York) 90(7), 6–12. 600 pp.
179

Nolan, C. (1993) The history of Falkland Islands trout. The Warrah: Stearley, R.L. (1992) Historical ecology of Salmoninae, with special
Newsl. Falklands Cons. 4, 3 pp. reference to Oncorhynchus. In: Mayden, R.L. (ed.), Systematics,
Nordeng, H. (1983) Solution to the ’char problem’ based on Artic Historical Ecology and North American Freshwater Fishes.
char (Salvelinus alpinus) in Norway. J. Fish. Res. Bd Can. 40, Stanford University Press, Stanford, pp. 622–758.
1372–1387. Stearley, R.F. and Smith, G.R. (1993) Phylogeny of the Pacific
Northcote, T.G. (1992) Migration and residency in stream trouts and salmons (Oncorhynchus) and genera of the family
salmonids – some ecological considerations and evolutionary Salmonidae. Trans. Am. Fish. Soc. 122, 1–33.
consequences. Nord. J. Freshw. Res. 67, 5–17. Tchernavin, V. (1939) The origin of salmon: Is its ancestry marine
Parnell, R. (183) The natural and economical history of the fishes of or fresh water? Salm. Trout Mag. 95, 120–140.
the river district of the Firth of Forth. Mem. Wernerian Nat. Hist. Thorpe, J.E. (1982) Migration in salmonids with special reference
Soc. 7, 161–460, 515–520. to juvenile movements in fresh water. In: Brannon, E.S. and Salo,
Pennant, T. (1776) Fishes. In: The British Zoology vol. 4. March, E.O. (eds.), Proceedings of the Salmon and Trout Migratory
for Cymmrodorian Society, London. Behaviour Symposium. School of Fisheries, University of Wash-
Paterson, A.M., Wallis, G.P. and Gray, R.D. (1995) Penguins, ington, Seattle, pp. 86–97.
petrels, and parsimony: Does cladistic analysis of behaviour Thorpe, J.E. (1987) Smolting versus residency: developmental
reflect seabird phylogeny? Evolution 49, 974–989. conflicts in salmonids. Am. Fish. Soc. Symp. 1, 224–252.
Regan, C.T. (1911) The Freshwater Fishes of the British Isles. Thorpe, J.E. (1988) Salmon migration. Sci. Progr. Oxford 72, 345–
Methuen, London, 287 pp. 370.
Regan, C.T. (1914) The systematic arrangement of the fishes of the Vladykov, V.D. (1963) A review of the salmonoid genera and their
family Salmonidae. Ann. Mag. Nat. Hist. 13 (ser 8), 405–408. broad geographical distribution. Trans. R. Soc. Can. 1, 459–505.
Rosen, D.E. (1974) Phylogeny and zoogeography of salmoniform Unwin, M.J., Kinnison, M.T. and Quinn, T.P. (1999) Exceptions
fishes and relationships of Lepidogalaxias salamandroides. Bull. to semelparity: Postmaturation survival, morphology and ener-
Am. Mus. Nat. Hist. 153, 265–326. getics of male chinook salmon (Oncorhynchus tshawytscha).
Rounsefell, G.A. (1958) Anadromy in North American Salmonidae. Can. J. Fish. Aquat. Sci. 56, 1172–1181.
Fish. Bull. U.S. Fish Wildl. Serv. 62, 171–185. Wood, C.C. and Foote, C.J. (1996) Evidence for sympatric diver-
Saunders, R.L., Henderson, E.B. and Glebe, B.D. (1982) Precocious gence of anadromous and non-anadromous morphs of sockeye
sexual maturation and smoltification in male Atlantic salmon salmon (Oncorchynchus nerka). Evolution 50, 1265–1279.
(Salmo salar). Aquaculture 28, 211–229.
Silverstein, J.T. and Hershberger, W.K. (1992) Precocious matur-
ation in coho salmon (Oncorhynchus kisutch): estimation of
additive genetic variance. J. Hered. 83, 282–286.

You might also like