Not be cited without prior reference to the authors
International Council for the
Exploration of the Sea
ICES CM 2002/ Y:20 (poster)
The Effects of Fishing on the Genetic Composition of Living
Marine Resources (Session Y)
Morphometric and meristic variation in Argentine hake (Merluccius hubbsi)
and southern hake (Merluccius australis) from the southwest Atlantic.
G.J. Pierce, M.B. Santos, A.J. Bishop, J.M. Bellido, M. Rasero and J.M. Portela
G.J. Pierce, M.B. Santos and A.J. Bishop: Department of Zoology, Aberdeen University,
Tillydrone Avenue, Aberdeen, AB24 2TZ, UK. [tel: +44 1224 272459, fax: +44 1224
272396, e-mail: g.j.pierce@abdn.ac.uk
J.M. Bellido, M. Rasero and J.M. Portela: Instituto Español de Oceanografía (IEO), P.O.
Box 1552, 36200, Vigo, SPAIN, [Tel: +34 986 492111, Fax: +34 986 492351, e-mail:
julio.portela@vi.ieo.es
Samples of Merluccius hubbsi (N=147) were collected from the southwest Atlantic, in the
high seas at 42o S and 46oS and from the waters around the Falkland Islands. A sample of M.
australis (N=23) was also collected in Falkland Islands waters. Variation in the population
structure was investigated using multivariate analysis of a total of external and skeletal
morphometric data, counts of fin rays and teeth, and measurements on scales. All
measurements were standardised to mean body size and each character set was analysed
separately. Principle components analysis and discriminant analysis were used to identify
differences between M. hubbsi from different areas and differences between the two species.
The results indicate the presence of two groups of Merluccius hubbsi within the study area,
one found on the high seas and one in the waters around the Falkland Islands. Results are
discussed in relation to the reproductive and trophic migrations of M. hubbsi, the relative
importance of genetic and environmental differences, and the results of similar studies from
the southwest Atlantic.
Keywords: South West Atlantic, hakes, stock discrimination
Introduction
The fishing grounds of the Patagonian Shelf support some of the most important fisheries in
the world, with hake (Merluccius hubbsi and Merluccius australis) and cephalopods (Illex
argentinus and Loligo gahi) being the main commercial species for fleets from coastal states,
EU and Far East countries. The great abundance of marine resources among parallels 35º and
54º South, is associated with the Subtropical Convergence formed by the Brazil and
Falkland/Malvinas currents. The mixing of the flow of La Plata river and the western branch
of the Falkland/Malvinas Current generates areas of high plankton production on the shelf
(Podestá, 1987).
The annual mean catch of the different fleets is around 600,000 tons of hake. These fleets
also catch important quantities of squid and accompanying species such as Patagonian
toothfish (Dissostichus eleginoides), Kingclip (Genypterus blacodes), Hoki (Macruronus
magellanicus), Red cod (Salilota australis), Southern blue whiting (Micromesistius
australis), etc.
Besides the vessels fishing in the EEZ's of Argentina and Uruguay, several hundred ships
with Spanish, British, Portuguese, Italian, Russian, Japanese, Korean or Taiwanese flags,
usually operate around the Falklands/Malvinas waters and in the high seas. The majority of
these ships are jiggers fishing for squid.
The fishing grounds of the Patagonian Shelf are currently the most important to the Spanish
bottom trawler freezing fleet, mainly based in Vigo (NW Spain). This fleet is composed of
about 40 vessels, besides another 20 that operate in joint ventures with Falkland flag. It is
estimated that this fleet generates approximately 2,000 direct offshore jobs, and more than
10,000 indirect onshore jobs. The value at first sale of the catches of the Spanish fleet in this
area is estimated at around 70,000 million pesetas per year (411 MEURO).
Until now, the only areas with regulation measures are the Argentinean and Uruguayan
EEZ's and inside the Falkland Islands Conservation Zone (FICZ) (Basson et al., 1996).
Partial stock assessments of hake have been made in Argentinean waters (Bezzi et al., 1994)
and around the Falkland Islands (Tingley et al., 1995). Although there is a bilateral
Argentine/UK SAFC trying to regulate exchange of fishery data between the two countries,
one of the difficulties for the assessment and management of these straddling stocks is the
lack of an international commission involving all the countries which prosecute the fisheries
for the assessment and discussion of management measures for the conservation of the stocks
(Martínez et al., 1997).
Merluccius hubbsi is a migratory species, spawning along the Atlantic coast of South
America and migrating both along the coast and into deeper waters linked to the Brazil/
Falklands confluence and areas of localised upwelling, where food is abundant. Spawning is
thought to take place in at least two areas, the Bonaerense spawning ground in Uruguayan
waters (autumn spawning) and off the coast of Argentina (summer spawning) (Otero et al.,
1986). The congeneric M. australis (also known as M. polylepis) is rarely caught in the high
seas north of the Falklands Islands but is a significant component of Falklands fishery
catches, although fishery statistics do not distinguish between the two species (Falkland
Islands Government, 2001).
The present paper describes work carried out as part of the CEC DG Fisheries Study Project
“Data Collection for Stock Assessment of Two Hakes (Merluccius hubbsi and M. australis)
in International and Falkland Waters of the SW Atlantic”. One aim of this project was to try
to confirm the number of stocks of M. hubbsi present in the SW Atlantic, information
required to underpin management decisions.
Multivariate analysis of morphometric and meristic characters is a standard tool for defining
population units and differentiating between genera, species, sub-species, and groups of
animals (e.g. Fridriksson, 1958; Thorpe, 1975; Boetius, 1980; Pierce et al, 1994a,b; Tudela,
1999, Bolles & Begg, 2000). Meristic characters are enumerable morphological features such
as fin rays, gill rakers and vertebrae, where as morphometric characters are those obtained by
measurements of body parts.
Morphometric and meristic differences can arise when genetic isolation allows genotypic
differences to arise through local differences in selective pressures, mutation or genetic drift
(Hadon & Willis, 1995). Phenotypic differences may also have an environmental basis, e.g.
local food and temperature regimes may influence growth patterns (Mamuris et al, 1998).
Previous studies on morphometric variability in M. hubbsi in Argentine/Uruguayan waters
indicated the presence of three possible stocks in the south Atlantic (Perrotta & Sánchez,
1992), the first north of 42oS in the Rio Plata region, the second between 44-48oS in the
Golfo san Matias region and the third south of 48oS around the Falkland islands and southern
Patagonia (spawning area unknown). However, an earlier study by Sardella (1984) proposed
the existence of two stocks, a northern stock was found north of 42oS and a southern stock
was south of 42oS.
In the present study therefore, our aim was to obtain further data on stock differentiation
within M. hubbsi from the high seas of the southwest Atlantic and from around the Falkland
islands. Samples of M. australis were also collected. The two Merluccius species not
distinguished in catches and it was thus also of interest to determine which morphometric
characters were reliable indicators of species identity.
Since results based on analysis of a single character set would require corroboration (Thorpe,
1975), external and skeletal morphometric characters, meristic characters (fin ray and tooth
counts), and measurements on scales were utilised to provide several independent character
sets.
Materials and Methods:
Whole hakes where collected from fishing vessels operating in the high seas of the Southwest
Atlantic and from around the Falkland Islands during 2000 and 2001 (see Table 1). Seven
groups of fish were available, each representing a particular sampling location and time. Five
groups were M. hubbsi from the “high seas”, one was M. hubbsi from south of the Falkland
Islands and the final sample was of M. australis. To minimise any effect of morphometric
differences between sexes, we focused on female hakes. However, 12 of the fish collected
turned out to be males. These were included in the analysis since no evidence of sex
differences in shape could be found. Of the females, the majority were found to be at
maturity stage 2, with only two mature fish (stages 3 and 4). Most of the males were stage 1
and none were mature. A major limitation in the analysis is that it was not possible to identify
the precise sampling location or date for the majority of the fish due to incomplete data
provision by observers.
The fish were stored frozen (-20oC) prior to analysis. Analysis took place at the Instituto
Español de Oceanográfia inVigo (Spain) and at the University of Aberdeen (UK). Particular
measurements were always made by the same person to minimise errors. Selection of
external morphometric and meristic characters was based on previous work (e.g. Sardella,
1984; Perrotta & Sánchez, 1992; Murta, 2000). Measurements on cranial bones were based
partly on Morales & Rosenlund (1979), referring also Mujib (1967). The following data sets
were collected for each fish (see Table 2):
a) 18 external morphometric measurements including total length (Fig. 1). Sample 4 was
not included since the fish thawed out in transit and arrived in poor condition.
b) 3 counts of fin rays
c) 2 measurements of scale dimensions (scales were taken above the lateral line near to
the operculum; up to 10 scales were measured per fish for several fish in each group).
d) 16 measurements on cranial bones. For paired bones, both bones were measured – in
the analysis either left or right measurements were used, depending on which set had
fewer missing values. Measurements were completed on the first 5 sets of samples.
e) 3 counts of teeth (premaxilla, dentary, vomer).
Data were screened for errors, using regression analysis (MINITAB statistical software) to
detect any outliers in relationships between each variable and body length. All values
differing by more than 3.5 standard residuals from the predicted value were checked against
original data sheets and transcription errors corrected. For measurements on bones, errors
were also checked against the stored skeletal material. Any remaining value differing by
more than 4 standard residuals from its predicted value was considered to be erroneous and
deleted from the analysis. Residuals from regressions were tested for departure from
normality. Although some statistically significant departures were observed, most
distributions of residuals were sufficiently close to normal that no transformations were
applied.
External morphometric data, skeletal measurements and tooth counts were standardised using
total length, thus normalising the fish in a sample to a single arbitrary size common to all
samples. A general linear modal was fitted to data for each particular measurement. The
standardised measurement Y' for each of the variables is then:

YY b X  X

where Y is the original observation, b is the regression slope and X is the total length.
Principle components analysis (PCA) was carried out separately on each character set. Plots
of second axis versus first versus scores were examined for evidence of segregation of fish
from different areas. Stepwise multiple regression was used to identify which variables
would be useful to discriminate between species and areas. Discriminant analysis (with crossvalidation) was then used to test how reliably the two species could be separated and, for M.
hubbsi, how reliably high seas fish could be distinguished from Falklands fish.
Results
External morphometrics
All samples except sample 4 (which had thawed in transit) were used in analysis of external
morphometric characters. An initial PCA on unstandardised data showed that the first
principle component explained almost 90% of the variability. This first axis was, as expected,
dominated by the effect of body size. Using standardised data, the first axis accounted for
only 23% of variation and eleven principal components were needed to explain 90% of
variation. There was good separation of M. australis from M. hubbsi along axis 1. There was
also good separation of M. hubbsi from the Falklands from those collected in the High Seas.
Excluding the four variables for which slopes of regressions on body length were nonhomogeneous, the first PCA axis explained 27% of variation and a plot of axis 1 and 2 scores
again showed separation of M. australis and M. hubbsi from the Falklands (Fig 2).
Stepwise multiple regression analysis suggested that 8 variables (FG, NO, PQ, QR, XX, AK,
YY and CD) would contribute towards species identification. Use of these variables resulted
in 98.5% successful species discrimination. Considering M. hubbsi alone, only three
variables (FG, XX, PQ) were selected by stepwise multiple regression as being useful to
discriminate area. Use of these resulted in 84% successful classification of M. hubbsi into
Falklands and High Seas groups.
Skeletal morphometric characters
Skeletal measurements were completed on the first five groups of fish. PCA on standardised
data showed some separation of both M. australis and the Falklands sample of M. hubbsi,
although the separation is less clear than for the external measurements. The first axis
explained approximately 38% of variation and 10 axes were required to explain 90% of
variation. Several of the variables showed non-homogeneous slopes. Excluding these, a
similar picture was obtained, although the new 1st PCA axis explained only 37% of variation.
Use of stepwise multiple regression suggested that species identification could be based most
reliably on opercular depth, the maximum distance between the two arms of the posttemporal and Premaxilla length. Discriminant analysis based on these standardised variables
resulted in 95% success in classifying fish to the correct species. Similarly, discriminant
analysis based on opercular length was 67% successful in distinguishing Falklands M. hubbsi
from high seas fish, although no other variables improved this discrimination.
Fin ray counts
The first PCA axis explained 70% of variation in the three external meristic variables. PCA
plots indicated that these variables provide a reliable way of distinguishing the two species,
but were not useful to distinguish between M. hubbsi from different areas. Discriminant
analysis confirmed that use of these characters would by 99% successful in identifying fish to
species but only 76% successful in discriminating between M. hubbsi from the high seas and
Falklands areas.
Tooth counts
The internal meristic variables proved to be dependent on body size and were therefore
standardised prior to PCA. The first PCA axis explained 71% of the variation in the three
original variables. Although there was considerable overlap, axis 1 scores for M. australis
were lower than for M. hubbsi. There were differences between some groups of M. hubbsi,
but Falklands M. hubbsi were not easily distinguished from high seas fish.
Discriminant analysis using these three characters was only 66% successful in identifying
fish to species and only 61% successful in distinguishing Falklands M. hubbsi from high seas
fish.
Combining fin ray and tooth data provided no improvement in discrimination results. Success
in identifying fish to area was 66%.
Scales
A total of 1213 scales was measured. Scale dimensions increased with fish size, although
slopes of scale length and width versus body length were non-homogeneous.
One-way analysis of variance showed that scales of M. hubbsi were significantly longer
(P=0.003) and wider (P<0.001) than those of M. australis. This was despite the fish in the M.
australis sample having a higher average body length than the M. hubbsi sampled (P<0.001).
The sample of M. hubbsi from the Falklands had higher average body length than samples
from the high seas (P<0.001) and also had longer and wider scales (P<0.001). ANCOVA
conformed that there were no significant between-area differences in scale size once body
length was taken into account.
Discussion
Multivariate analysis of both external and internal morphological characters has been used
successfully on other hake populations and other fish species as a tool for separating groups
from distinct geographical regions and also for differentiating between stocks.
In the present study, analysis of external morphometric variation in Merluccius hubbsi taken
from the high seas and from around the Falkland Islands indicated the presence of two
stocks, one located on the high seas and one around the waters of the Falkland Islands.
Skeletal characters proved less useful in discriminating area.
The study also identified several physical characteristics that can be used to distinguish M.
hubbsi from M. australis.
Torres et al. (2000) used sagittal otolith shape and size as a tool for stock discrimination in
M. hubbsi, as well as M. gahi from Chile and Peru and M. merluccius from the
Mediterranean and the North Atlantic. In common with this study, they reported only two
distinct groups of M. hubbsi.
Perrotta & Sánchez (1992) identified three M. hubbsi
populations on the Argentinean and Uruguayan shelf, two of which were located to the north
of Argentina. However, samples taken from similar areas by Torres et al. (2000) showed
little morphological variation, indicating the presence of just one geographical population
north of Argentina.
The differences in results between these studies may be a result of the reproductive and
trophic seasonal migrations of this species, with fish from different (reproductively isolated)
stocks overlapping in distribution outside the breeding season. Trophic studies show that M.
hubbsi migrate from the coast into deeper water (high seas) to feed on abundant squid and
fish species found there.
The present study was based on the fishery in the High Seas and off the Falklands, which
takes hake mainly outside the spawning season. The fish examined in the present study were,
consequently, mainly immature females. Thus some stock differences may have gone
undetected due to samples containing fish from more than one breeding group.
Ideally, to identify separate breeding stocks, it would be useful to sample mature fish on the
spawning grounds. Mature M. hubbsi can be found all year round, with two peaks of
breeding activity: in the austral winter (May to July) at the northern end of their range (3538o S) and in the austral summer (October-March) on the Patagonian Shelf (Cousseau &
Perrotta, 2000).
The third population identified by Perrotta & Sánchez (1992) was located in the Patagonian
region of Argentina. M. hubbsi are known to undergo trophic migrations from the coast of
Argentina to the waters around the Falklands Islands, which could indicate that the group
from around the Falkland Islands identified in the present study is the same as that from
Patagonian waters identified by Perrotta & Sánchez.
The extent to which observed geographical differences in morphometric characteristics of M.
hubbsi have a genetic or environmental basis remains an important question. It is interesting
to note that, for several of the external morphometric characters which show significant
variation, values for M. australis and for Falklands M. hubbsi tend to be at opposite extremes,
e.g. standardised pre-orbital length is higher in M. australis than M. hubbsi, but is lower in
Falklands M. hubbsi than in the high seas fish. This could indicate the selective effect of
interspecific competition, since it is only around the Falkland islands that these two species
overlap substantially in their range. Thus, Falklands M. hubbsi tend to be more different from
M. australis than are high seas M. hubbsi.
The results on species discrimination from the present study are consistent with those of by
Cousseau & Cotrina (1981), who found the scales to be smaller in M. australis than in M.
hubbsi and also reported that the colour is dark grey in M. australis compared to silver grey
in M. hubbsi. Analysis of fin ray counts showed that M. australis have more fin rays than M.
hubbsi.
Cousseau & Cotrina (1981) identified several differences in the morphometric characters of
M. australis and M. hubbsi. Differences are also summarised by Cousseau & Perrotta (2000):
M. australis has smaller eyes and a longer snout, reaches a larger body size and has more
than 40 fin rays in the second dorsal fin. Generally however, M. australis and M. hubbsi are
similar in appearance, and also closely resemble M. gahi found on the Chilean/Peru coast
(Roldan et al., 1999). This reflects underlying genetic similarity; both M. australis and M.
hubbsi are clustered along with other American species (Roldan et al., 1999).
Further research is still needed to confirm the number of stocks of hakes within the
Southwest Atlantic. This does not simply require wider simultaneous sampling for many
locations, although this is desirable. Political considerations have so far restricted the scope
of studies, and work is needed which encompasses both Argentinean and Falklands territorial
waters as well as the adjacent high seas. This information is required for reliable assessment
of stocks to support long-term sustainability in the fishery.
Acknowledgements
This work was supported by CEC DG Fisheries under Study Project No 99/016, Data
Collection for Stock Assessment of Two Hakes (Merluccius hubbsi and M. australis) in
International and Falkland Waters of the SW Atlantic.
References
Bolles, K.L. & Begg, G.A., 2000. Distinction between silver hake (Merluccius bilinearis)
stocks in US waters of the northwest Atlantic based on whole otolith morphometrics.
Fishery Bulletin 98, 451-462.
Boetius, J., 1980. Atlantic Anguilla. A presentation of old and new data of total numbers of
vertebrae with special reference to the occurrence of Anguilla rostrata in Europe. Dana 1,
1-28.
Cousseau, M.B. & Cotrina, C.P., 1981. Observaciones sobre diferencias morfologicas entre
la merluza comun (Merluccius hubbsi) y la merluza austral (Merluccius polylepis). Rev.
Inv. Des. Pesq. 2, 47-56
Cousseau, M.B. & Perrotta, R.G., 2000. Peces marinos de Argentina: biología, distribución,
pesca. INIDEP, Mar del Plata, 167 pp.
Falklands Islands Government, 2001. Fisheries Department Fisheries Statistics, Volume 6.
FIG Fisheries Department, Stanley, 69 pp.
Fridriksson, A., 1958. The tribes of North Coast herring of Iceland with special reference to
the period 1948-1955. Rapports du Conseil pour l’Exploration de la Mer 143, 36-44.
Haddon, M. & Willis, T.J., 1995. Morphometric and meristic comparison of orange roughy
(Haplostethus atlanticus: Trachichthyidae) from the Puysegur Bank and Lord Howe Rise,
New Zealand, and its implications for stock structure. Marine Biology 123, 19-27.
Mamuris, Z., Apostolidis, A.P., Panagiotaki, P., Theodorou, A.J. & Triantaphyllidis, C.,
1998. Morphological variation between red mullet populations in Greece. Journal of Fish
Biology 52, 107-117.
Morales, A. & Rosenlund, K., 1979. Fish bone measurements. Steenstrupia, Zoological
Museum, Copenhagen.
Mujib, K.A., 1967. The Cranial Osteology of the Gadidae. Journal of the Fisheries Research
Board of Canada 24, 1315-1375.
Murta, A.G., 200. Morphological variation of horse mackerel (Trachurus trachurus) in the
Iberian and North African Atlantic: implications for stock identification. ICES Journal of
Marine Science 57, 1240-1248.
Otero, H.O., Giangiobbe, M.S. & Renzi, M.A., 1986. Aspecto de la estructura de la merluza
común (Merluccius hubbsi). II. Distribución de tallas y edades, estadios sexuales,
variaciones estacionales. Publicaciones del Comité Técnico Mixto del Frente Marítimo 1,
147-179.
Perrotta, R.G. & Sánchez, F., 1992. Una vez más acerca de la identificación de las
poblaciones de merluza (Merluccius hubbsi) en la plataforma argentino-uruguaya. Frente
Marítimo 12, 39-46.
Pierce, G.J., Hastie, L.C., Guerra, A., Thorpe, R.S., Howard, F.G. & Boyle, P.R., 1994.
Morphometric variation in Loligo forbesi and Loligo vulgaris: regional, seasonal, sex,
maturity and worker differences. Fisheries Research 21, 127-148.
Pierce, G.J., Thorpe, R.S., Hastie, L.C., Brierley, A.S., Boyle, P.R., Guerra, A., Jamieson, R.
& Avila, P., 1994. Geographic variation in Loligo forbesi in the Northeast Atlantic:
analysis of morphometric data and tests of causal hypotheses. Marine Biology 119, 541548.
Roldán, M. I., García-Marín, J. L., Utter, F. M. & Pla, C., 1999. Genetic relationships among
Merluccius species. Heredity 83, 79-86.
Sardella, N.H., 1984. Mixosporidios parásitos musculares de peces del Mar Argentino
(incidencias, reacciones de respuesta ante la agresión parasitaria, consideraciones
zoogeográficas y aspectos tecnológicos). PhD thesis, Universidad Nacional de la Plata,
Argentina.
Thorpe, R.S., 1976. Biometric analysis of geographic variation and racial affinities.
Biological Review 51, 407-452.
Torres, G.J., Lombarte, A. & Morales-Nin, B., 2000. Sagittal otolith size and shape
variability to identify geographical intraspecific differences in three species of the genus
Merluccius. Journal of the Marine Biological Association of the United Kingdom 80, 333342.
Tudela, S., 1999. Morphological variability in a Mediterranean, genetically homogeneous
population of the European anchovy, Engraulis encrasicolus. Fisheries Research 42, 229243.
Table 1 Summary of fish sampled: area, species, length (mean and range, cm),
number sampled
Sample
Area
Species
1
2
3
4
5
6
7
High Seas
Falklands
High Seas
High Seas
Falklands
High Seas
High Seas
M. hubbsi
M. australis
M. hubbsi
M. hubbsi
M. hubbsi
M. hubbsi
M. hubbsi
Length (cm)
Mean and range
54.9 (40.0-67.0)
69.4 (51.5-79.0)
51.0 (35.0-75.0)
56.9 (36.0-76.5)
70.6 (45.5-89.0)
59.3 (32.5-90.0)
50.8 (31.5-63.5)
Number
sampled
33
23
30
27
23
15
19
170
Table 2. Morphometric and meristic data collected on M. hubbsi and M. australis.
Codes for external morphometric measurements refer to Figure 1.
Variable
Code
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
AM
AK
AE
NO
CD
OP
PQ
AF
NS
NQ
QR
FG
HI
ST
AB
XX
ZZ
YY
19
20
SL
SW
21
22
23
24
25
26
27
28
29
30
31
32
33
DLL
DH
DUL
ML
MH
PML
PMNH
VW
CW
CH
PTLL
PTSL
PTD
34
35
36
OSL
OCL
OD
1
2
3
FRA
FRB
FRC
4
5
6
DTOO
PTOO
VTOO
Measurement (cm)
External morphometrics
Total length
Standard (pre-caudal) length
Head length
Pre-orbital length
Eye diameter (pupil)
Orbital diameter
Post-orbital length
Pre-dorsal length
Pre-anal length
Pre-pectoral length
Pectoral fin length
Length of 1st dorsal fin
Length of 2nd dorsal fin
Length of anal fin
Length of mouth
Body height
Height of caudal peduncle
Body width
Scale size
Scale length
Scale width
Internal morphometrics
Dentary lower length
Dentary height
Dentary upper length
Maxilla length
Maxilla height
Premaxilla length
Premaxilla nose height
Vomer width
Cranium width
Cranium height
Post-temporal length 1
Post-temporal length 2
Distance between post-temporal
arms
Opercular length 1
Opercular length 2
Opercular depth
External meristics
Number of 1st dorsal fin rays
Number of 2nd dorsal fin rays
Number of anal fin rays
Internal meristics
Dentary tooth count
Pre-maxilla tooth count
Vomer tooth count
Figure 1. External morphometric variables for hake. Each measurement in Table 2 is defined as the distance between two points on the above diagram
(from Perrotta & Sánchez, 1992).
5
4
3
PCA axis 2
M.h. High Seas 1
2
M.h. High Seas 3
1
M.h. High Seas 6
M.h. High Seas 7
0
M.h. Falklands
-1
M.a. Falklands
-2
-3
-4
-5
-6
-4
-2
0
2
4
6
PCA axis 1
Figure 2. Plot: scores on principal component axes 1 and 2, both species, all areas (using standardised external morphometric data, excluding variables
for which slopes of relationships with body length were non-homogeneous. Identities of groups are as in Table 1. Group 2 (M. australis) are seen to be
quite distinct and there is some separation also of group 5 (M. hubbsi from the Falklands).