Geography Reference
In-Depth Information
network measurements. Meristic
4
features can also differ
between stocks. For instance, Claytor and MacCrimmon
(1988) visually counted vertebra and dorsal and anal fin
rays from radiographs in order to differentiate regional
stocks of Atlantic salmon (
Salmo salar
).
As for species recognition, subtle variations in size and
contour of otoliths were used in the discrimination of
individual populations/stocks (e.g. Campana and Cas-
selman, 1993) and in the determination of phylogenetic
lines (e.g. Gaemers, 1984). Furthermore, fish stock iden-
tification can rely on scale pattern analysis. Fish otoliths
and scales grow by accretion, as more bone is period-
ically added along their periphery. Subsequently, their
diameter increase reflects body growth and circuli
5
are
closer together during periods of slow growth, such as
winter in temperate regions. This growth pattern creates
alternative dark and light bands (called annuli), which
can be counted and analysed through IA. For diadromous
species, the freshwater-ocean transition is usually asso-
ciated with a change from thin, narrowly spaced circuli
to thick, widely spaced ones (Bernard and Myers, 1996).
Moreover, the freshwater and early marine portions of
the scale were shown to differ respect to the origin of
fish (i.e. hatchery
vs
wild specimens, see Davis and Light
1985). Using acetate impressions of scales and subsequent
IA, Bernard and Myers (1996) showed that hatchery steel-
head salmon (
Oncorhynchus mykiss
) had larger freshwater
zones and more freshwater circuli than wild specimens.
They conclude that this technique has good potential for
estimating proportions of hatchery and wild steelhead in
high-seas mixed-populations.
Farmed and wild fish may also differ in their general
body shape. Using geometric morphometrics methods
(Procrustes coordinates of landmarks and visualisation
by thin-plate splines), captively reared adults were dif-
ferentiated from wild ones by sharply reduced sexual
dimorphism as well as numerous differences in body
shape (Hard et al., 2000).
As for species recognition, colour features are less
used than shape ones for stocks differentiation. How-
ever, Strachan and Kell (1995) used ten shape features
and 114 colour features to discriminate between haddock
(
Melanogrammus aeglefinus
) stocks from two different
fishing regions, demonstrating that IA methods are effi-
cient to segment fish images from colour information.
17.2.1.3 Sexual dimorphism
Sexual dimorphism is common in animals where males
and females have distinct roles in mating and courtship.
Aquatic species are not an exception to this rule, and such
a dimorphism can be found in a wide variety of freshwater
taxa. Using Fourier analysis of outlines, Bertin et al. (2002)
showed that three characters (pleotelson, paraeopods 4
and 5) differed significantly in shape between males and
females of
Asellus aquaticus
(Crustacea). By comparing
three species in the genus
Poecilia
(Pisces), Ptacek (1998)
showed that differences in behaviour and morphometrics
in males could play a role in female mate choice. There-
fore, IA can rapidly sort individuals by gender, through
morphological differences among sexes. For that purpose,
Zion et al. (2008) used algorithms derived from shape (i.e.
locating landmark positions on fish contours and extrac-
tion of shape-related features) and colour differences
between female and male guppies (
Poecilia reticulata
)
to classify them. Identification accuracy was approxi-
mately 90% using shape features, approximately 96%
using colour features and was slightly improved when
both colour and shape features were used. In a same way,
studying skin colour in newts (
Notophthalmus viridescens
viridescens
) through IA methods, Davis and Grayson
(2007) showed that males were statistically greener than
females, although this effect depended upon life-stage. If
males and females often differ in external shape or colour
features, sexual dimorphism can also be detected through
analysis of histological images. Under a light microscope
and using IA software, Hagen et al. (2006) showed that the
total number of fast muscle fibres per trunk cross section
was higher in females than males prior to sexual matura-
tion in Atlantic halibut (
Hippoglossus hippoglossus
). These
results illustrated a sexual dimorphism of muscle fibre
recruitment patterns in some fish species.
17.2.2 Characterisationof life-history traits
andontogeneticstages
Characterisation of life-history traits is a very important
task in population biology and evolution. For example,
estimation of individual growth and reproductive invest-
ment is central in many studies (e.g. Jennings and Philipp,
1992). Moreover, despite the idiosyncratic nature of many
invasions (Marchetti et al., 2004), it has been hypothe-
sised that life-history traits such as early maturity, high
fecundity and asexual reproduction are associated with
successful invading species (Lodge, 1993). Therefore,
numerous studies have characterised various life-history
traits, sometimes using IA methods.
4
Meristic features are all the variables that can be counted on a
fish (e.g. number of fin rays, of vertebrae, of scales on the lateral
line, etc.).
5
Circuli are layers of roughly concentric circles of bone that appear
on fish scales and otoliths.
Search WWH ::
Custom Search