Allozymic Polymorphism and Phylogeny of the Family Channichthyidae
307
Heterozygosity in the Polar Environment
In marine fish, the mean heterozygosity is about 5-6% [17]. In our
sample, except for Chionodraco myersi, whose heterozygosity is 10%,
most species present a low polymorphism, but greater than that of the two
species analyzed by the same markers by Duhamel et al. [6]. McDonald et
al. [5], with seven species belonging to the Trematomus/Pagothenia
complex, indicated that multilocus heterozygosity varied from 1.8 to 7.0%
across taxa.
However, the data given here should be interpreted with caution:
Chionodraco myersi, with a 65 individuals sample, is the best sampled
species and the most polymorphic. The estimation of polymorphism level
for species represented by small samples is thus doubtful.
Phylogeny of the Channichthyidae
The Dollo parsimony tree (Fig. 2) can be considered as reliable since its
RI is relatively high (0.757). The most parsimonious tree maximizes
synapomorphy hypotheses and minimizes the ad hoc hypotheses of
homoplasy [18,19]. This rather high RI tends to show a relatively good
level of structured variability in these allozymic data, an observation that
is to be balanced with relatively low bootstrap proportions (below 70%).
Allozymic and morphological data provide congruent points and
incongruent ones. Comparing Iwami's [3] tree (Fig. 3) and our Dolloparsimony tree (Fig. 2), the congruent points are the basal position of
Champsocephalus; then, the emergence of Neopagetopsis and Pagetopsis
and the terminal position of the clade (Chaenocephalus, Cryodraco).
The two approaches disagree on the position of Channichthys, and the
relative positions of Chionodraco and Chaenodraco (monophyly of the
two in Iwami's [3] tree and paraphyly in our trees). Furthermore, the tree
of Iwami [3] is not completely congruent with the phenogram of
Voskoboinikova [4]; (see the tree Fig. 3, rerooted on Champsocephalus
for comparison). Voskoboinikova's tree differs on the positions of
Chaenocephalus, without being more congruent with ours.
However, Voskoboinikova's tree (originally rooted on Cryodraco
without justification) is not a phylogenetic tree sensu stricto but an
analysis of global similarity in developmental features. The comparison is
therefore difficult to make.
Our hypothesis concerning the paraphyly of Chionodraco does not
appear elsewhere. That paraphyly, whatever its robustness, may be due to
different sample sizes in Chionodraco and other genera.
307
Heterozygosity in the Polar Environment
In marine fish, the mean heterozygosity is about 5-6% [17]. In our
sample, except for Chionodraco myersi, whose heterozygosity is 10%,
most species present a low polymorphism, but greater than that of the two
species analyzed by the same markers by Duhamel et al. [6]. McDonald et
al. [5], with seven species belonging to the Trematomus/Pagothenia
complex, indicated that multilocus heterozygosity varied from 1.8 to 7.0%
across taxa.
However, the data given here should be interpreted with caution:
Chionodraco myersi, with a 65 individuals sample, is the best sampled
species and the most polymorphic. The estimation of polymorphism level
for species represented by small samples is thus doubtful.
Phylogeny of the Channichthyidae
The Dollo parsimony tree (Fig. 2) can be considered as reliable since its
RI is relatively high (0.757). The most parsimonious tree maximizes
synapomorphy hypotheses and minimizes the ad hoc hypotheses of
homoplasy [18,19]. This rather high RI tends to show a relatively good
level of structured variability in these allozymic data, an observation that
is to be balanced with relatively low bootstrap proportions (below 70%).
Allozymic and morphological data provide congruent points and
incongruent ones. Comparing Iwami's [3] tree (Fig. 3) and our Dolloparsimony tree (Fig. 2), the congruent points are the basal position of
Champsocephalus; then, the emergence of Neopagetopsis and Pagetopsis
and the terminal position of the clade (Chaenocephalus, Cryodraco).
The two approaches disagree on the position of Channichthys, and the
relative positions of Chionodraco and Chaenodraco (monophyly of the
two in Iwami's [3] tree and paraphyly in our trees). Furthermore, the tree
of Iwami [3] is not completely congruent with the phenogram of
Voskoboinikova [4]; (see the tree Fig. 3, rerooted on Champsocephalus
for comparison). Voskoboinikova's tree differs on the positions of
Chaenocephalus, without being more congruent with ours.
However, Voskoboinikova's tree (originally rooted on Cryodraco
without justification) is not a phylogenetic tree sensu stricto but an
analysis of global similarity in developmental features. The comparison is
therefore difficult to make.
Our hypothesis concerning the paraphyly of Chionodraco does not
appear elsewhere. That paraphyly, whatever its robustness, may be due to
different sample sizes in Chionodraco and other genera.