pusch
ARTICLE IN PRESS
Deep-Sea Research I 51 (2004) 1685–1708
www.elsevier.com/locate/dsr
Community structure and feeding ecology of mesopelagic
fishes in the slope waters of King George Island
(South Shetland Islands, Antarctica)
C. Puscha,Ã, P.A. Hulleyb, K.-H. Kockc
a
Alfred Wegener Institute for Polar and Marine Research, P.O. Box 120161, Columbusstrasse, 27568 Bremerhaven, Germany
b
Iziko Museums of Cape Town, P.O. Box 61, 8000 Cape Town, Republic of South Africa
c
Sea Fisheries Research Institute, Palmaille 9, 22767 Hamburg, Germany
Received 17 September 2003; received in revised form 18 June 2004; accepted 22 June 2004
Available online 24 August 2004
Abstract
The role of mesopelagic fishes in the Southern Ocean ecosystem and more particular their trophic effect on the
standing stock of mesozooplankton is at present poorly understood. To get a deeper insight in the Antarctic mid-water
ecosystem the mesopelagic fish community of the King George Island slope (South Shetland Islands) was sampled with
a pelagic trawl in 1996. The community structure was analysed and the feeding ecology was studied of the five most
abundant species. A total of 18 mesopelagic fish species in 10 families was identified. Of these, the Myctophidae was the
most important family by species number (9 species), individual number (98.5% of all individuals) and fish wet weight
(87.3% of the total weight). The assemblage was numerically dominated by four myctophids (Electrona antarctica,
Gymnoscopelus braueri, Gymnoscopelus nicholsi, Protomyctophum bolini) and one gempilyd (Paradiplospinus gracilis).
Multivariate statistical analysis of the mesopelagic fish data reveals two major groups of stations according to the
sampled depth: a shallow group of stations (295–450 m depth) and a deeper group of stations (440–825 m depth). The
change in relative abundance of mesopelagic fish species at 440–450 m coincides with the presence of warmer and denser
Circumpolar Deep Water at and below these depths. Deeper stations were characterized by a higher density and
increased diversity of mesopelagic fish species. The community patterns identified correlated well with the vertical depth
distribution of the most abundant species. Dietary analysis reveals that myctophids are mostly zooplanktivorous, while
the gempilyd P. gracilis is classified as a piscivorous predator. The small P. bolini feed mainly on copepods of the species
Metridia gerlachei, while the most important prey item of the larger myctophids E. antarctica, G. braueri, and G. nicholsi
were various species of euphausiids. Investigation of feeding chronology showed that G. nicholsi and P. bolini were
feeding day and night. Daily ration estimates for myctophid species ranged from 0.28% to 3.3% of dry body weight
(0.5–5.94% of wet body weight). Krill (Euphausia superba) were the most important food of E. antarctica and G.
nicholsi, accounting for 53.1% and 58.3% of the total food weight, respectively. The annual removal from the krill
ÃCorresponding author. Tel.: +49-471-4831-1652; fax:+49-471-4831-1425.
E-mail address: cpusch@awi-bremerhaven.de (C. Pusch).
0967-0637/$ - see front matter r 2004 Elsevier Ltd. All rights reserved.
doi:10.1016/j.dsr.2004.06.008
ARTICLE IN PRESS
C. Pusch et al. / Deep-Sea Research I 51 (2004) 1685–1708
1686
stock by both species was estimated to amount to 11.1–26.7% in the South Shetland Islands region. This estimate
emphasizes the important role of mesopelagic fish in the Antarctic ecosystem as a prevalent consumer of krill.
r 2004 Elsevier Ltd. All rights reserved.
Keywords: Marine fish; Myctophidae; Mesopelagic zone; Community structure; Stomach content; Predation; Daily ration
world oceans, they occupy the third level and are
1. Introduction
consumers of the second order. They are an
One-quarter of all known fish species in the important food source for the predators of higher
Southern Ocean live in the mesopelagic and trophic levels like benthopelagic fish (Bulman et
bathypelagic zones (Kock, 1992). Myctophids are al., 2002), seabirds (Guinet et al., 1996), fur seals
the dominant fish family in these zones, as far as (Cherel et al., 1997) and squid (Rodhouse et al.,
diversity, biomass and abundance are concerned. 1992; Phillips et al., 2001). Any estimation of
Thirty-three myctophid species are known from energy transport within the pelagic system must
the Southern Ocean, of which 11 have a circum- include analyses of the individual diet composition
polar distribution. Although their geographical of the mesopelagic fish and their rates of food
distribution and taxonomy have been described consumption.
(Hulley, 1981; McGinnis, 1982), comparatively Although the diet composition of the most
few studies have examined the vertical distribution abundant myctophid species is documented (Ro-
of oceanic micronekton by intensive depth strati- wedder, 1979a; Ascenio and Moreno, 1984;
fied sampling (Torres and Somero, 1988; Lancraft Williams, 1985; Kozlov and Tarverdiyeva, 1989;
et al., 1989; Piatkowski et al., 1994; Duhamel, Lancraft et al., 1991; Hoddell et al., 2000) only a
1998). All these studies indicate that the common few studies have investigated feeding chronology
Antarctic myctophids are diel vertical migrators. (Rowedder, 1979a) and estimated daily rations
King George Island is located in the southern (Gerasimova, 1990; Pakhomov et al., 1996).
Krill (Euphausia superba) plays an important
part of the Drake Passage and is strongly
influenced by the Antarctic Circumpolar Current role as the key prey item of a number of top
(ACC). The ACC is the major oceanographic predators, especially in the Atlantic Sector of the
feature of the Southern Ocean; it is an extensive Southern Ocean (Barlow et al., 2002). Because of
eastward flowing circumpolar current (Hofmann their high biomass (the total stock of the Southern
et al., 1996). The upper waters of the ACC in the Ocean was estimated by Lancraft et al. (1989) to
study area comprised Antarctic Surface Water account for 133–191 million tonnes), mesopelagic
(ASW) and the associated Circumpolar Deep fish could be one of the most important predators
Water (CDW), which flows from the Bellingshau- of oceanic zooplankton (Lancraft et al., 1989;
sen Sea into southern Drake Passage (Stein and Pakhomov et al., 1996). Numerous studies have
Heywood, 1994). The study site on the slope of shown that myctophids play a significant role in
King George Island is characterized by a shelf- the consumption of juvenile and adult krill
break front resulting in enhanced production and (Rembiszewski et al., 1978; Rowedder, 1979a;
a higher krill abundance compared to oceanic Williams, 1985). This conclusion has more recently
waters. For this reason the area north of King been challenged by a suggestion that a substantial
George Island is one of the most important krill consumption of krill occurs only during certain
fishing regions of the Southern Ocean (Ichii et al., periods and within specific regions (Pakhomov et
1996). al., 1996).
Myctophids play a significant role as consumers Finally, it should be noted that a preliminary
of zooplankton in the food web of the Southern analysis of community structure of the mesopela-
Ocean (Lancraft et al., 1989). As in the other gic fish assemblage from cruise ANT XIV/2 of RV
ARTICLE IN PRESS
C. Pusch et al. / Deep-Sea Research I 51 (2004) 1685–1708 1687
‘‘Polarstern’’ has already been presented by Hulley The sampling program consisted of 16 hauls,
et al. (1998), and the results used for a cladistic with the objective to sample three different depth
analysis of the myctophid tribe Electronini (Hul- horizons: 200–300 m, 400–500 m and near-bottom
ley, 1998). In the present study, however, commu- (2–28 m above seafloor) at soundings of 400, 600
nity structure of the mesopelagic community over and 800 m (Table 1) (Kock et al., 1998). Mesope-
the slope of King George Island is analysed in lagic fish were sampled with a pelagic trawl PT-
1088 with an estimated mouth opening of 200 m2
greater detail. The community pattern and the
vertical distribution are related to the feeding (width 20 m and a height of 10–12 m). The mesh
ecology of the most abundant mesopelagic fish size was 12 mm in the cod end. It was expected that
species. Four myctophids and one gempylid are juvenile myctophids (SLo30 mm) would not be
analysed with respect to diet, feeding chronology sampled adequately with this net configuration
and daily rations. In conclusion, an estimation is (Gartner et al., 1988). An SCANMAR depth
made of the predation impact of the mesopelagic sensor controlled the sample depth and net open-
fish community on the krill stock in the region of ing during trawling. Towing time varied between
King George Island. 30 and 60 min; trawl speeds ranged from 3.5 to
4.0 knots. Ship speed was increased during net
deployment and decreased during retrieval. This
procedure minimizes the effects of net contamina-
2. Material and methods
tion by fish resident in water layers above the
fishing depth. Station 73 was excluded from
2.1. Sampling
community analysis as the net snagged on the
bottom during trawling (Hulley et al., 1998). All
Data were collected during the cruise ANT
fishes were identified to species according to the
XIV/2 of RV ‘‘Polarstern’’ in November/Decem-
most recent keys (Gon and Heemstra, 1990). Fish
ber 1996. The study area was located over the
from the entire sample or a sub-sample of each
slope northwest of King George Island (South
species from each station were counted and
Shetland Islands), in southern Drake Passage
weighed, and standard lengths (SL) were taken
(Fig. 1).
˚S 1500 m
1000 m
61.5 500 m
51
73
70
57
52
53 68
55 69
59
58
63
64
65 54
60
200 m
orge Is.
62.0 King Ge
°W
60.0 59.5 59.0 58.5 58.0 57.5
Fig. 1. Sampling localities in 1996. Line indicates hydrographic section through the study area.
ARTICLE IN PRESS
C. Pusch et al. / Deep-Sea Research I 51 (2004) 1685–1708
1688
Table 1
Station data for 16 PT-1088 trawl samples
ST Sampling Locality Date Time Day Global Sample Bottom Ship
Timea
Longitude Latitude 1996 (Local) Radiation (wm2) Depth (m) Depth (m) Speed (km)
581430 W 611360 S
51 30 November 16:49–17:49 Day 197.8 400–450 786–875 4.0
591170 W 611450 S
52 1 December 08:00–08:30 Day 75.0 415–515 660–730 4.0
591190 W 611450 S
53 1 December 10:57–11:27 Day 153.0 273–302 635–738 4.0
591310 W 611480 S
54 1 December 14:50–15:20 Day 230.3 283–295 475–652 4.0
591340 W 611470 S
55 1 December 17:35–18:05 Day 145.8 400–450 680–995 4.0
591160 W 611430 S
57 3 December 08:19–08:50 Day 179.7 397–465 730–790 4.0
591130 W 611470 S
58 3 December 20:26–21:26 Day 12.4 283–325 380–387 3.5
591320 W 611480 S
59 3 December 23:20–00:20 Night 0.0 431–495 850–1058 3.7
591440 W 611480 S
60 4 December 02:15–03:18 Night 0.0 380–440 681–690 3.5
591320 W 611480 S
63 4 December 20:30–21:11 Day 98.5 520–580 555–608 4.0
591350 W 611470 S
64 5 December 01:06–01:41 Night 0.0 610–640 731–792 4.0
591510 W 611490 S
65 5 December 04:26–04:56 Day 80.6 750–800 1287–1468 4.0
591090 W 611440 S
68 5 December 21:10–21:41 Day 46.0 340–360 360–367 3.5
591170 W 611460 S
69 6 December 01:01–01:32 Night 0.0 560–597 584–599 3.7
591020 W 611380 S
70 6 December 04:50–05:20 Day 23.0 790–825 810–833 3.7
581440 W 611360 S
73 6 December 21:07–21:10 Day 20.0 550–575 570–580 3.5
a
Defined by sunset 22:06 and sunrise 03:23.
to the nearest millimetre with sliding callipers. (DWi ) were calculated (George and Hadley, 1979).
Hydrographic data were collected by conductivity, By the following equation, all three indices were
temperature and depth casts (CTD, 22 stations). combined to describe the prey utilization by the
The CTD was deployed at each station in advance ‘Relative Importance Index’ (RI) for each prey
category i (George and Hadley, 1979; Hyslop,
of the trawl tows.
1980):
2.2. Diet analysis ðFi þ Ni þ DWiÞ Â 100
RI i ¼ Ps ; ð1Þ
i¼1 ðFi þ Ni þ DWiÞ
Diet analysis was performed on the five most
abundant mesopelagic fish species. A maximum of
where s is the number of prey categories.
20 individuals of these species was selected from
Feeding chronology was analysed by the Sto-
each sample. In samples containing 420 speci-
mach Content Index (SCI):
mens individuals were chosen haphazardly. Fish
dry weight of stomach content
were weighed wet, measured (SL, mm below) and SCIð%Þ ¼ Â 100:
body dry weight
the whole stomach removed. The dry weight of fish
specimens was determined by oven-drying speci- ð2Þ
mens at 80 1C until constant weight was reached.
Prey organisms were identified to the lowest In addition, the stage of digestion of each prey
possible taxon and measured under a binocular item was determined by the modified method of
microscope. Dry weight of the prey organisms was Pearcy et al. (1979): Stage 1=undigested prey,
reconstructed by length–weight regressions taken Stage 2=slightly digested with some appendages
from the literature (Mizdalski, 1988; Groeger damaged, but body shape still preserved, Stage 3:
et al., 2000). body shape of prey deformed. The ratio of
Three indices, the frequency of occurrence (F i ) digestion stages was calculated for each time
of each prey item in non-empty stomachs, the interval based on these criteria.
Daily ration (mean daily food consumption, C w )
percentage of each food item by number (N i ) to
the total number, and the percentage by dry weight of the four myctophids was investigated by the
ARTICLE IN PRESS
C. Pusch et al. / Deep-Sea Research I 51 (2004) 1685–1708 1689
method of Eggers (1977): intensity were log transformed to validate the use
of normalized Euclidean distance for the calcula-
C w ¼ I Â 24=T; ð3Þ
tion of the abiotic similarity matrix. We applied
where I is the daily average SCI (%) and T the gut the Spearman Rank correlation to relate the biotic
passage time (h). matrix based on mesopelagic fish abundances with
In this study, no gut passage time data were the abiotic similarity matrix (Clarke and Ains-
collected. Therefore, data from the literature were worth, 1993).
used. Two estimates of egestion times for Antarc- Two different water masses were discernible
tic myctophids are available: the first estimate by over the slope of King George Island, separated at
Rowedder (1979a) for Electrona antarctica and the a depth of 450 m (see Section 3). Mesopelagic fish
second by Gerasimova (1990) for E. carlsbergi. samples above and below this depth were com-
These studies estimated the egestion time to be 8 pared with different sub-routines of the Primer
and 8.5 h, respectively. In our study, we used the computer program.
8.5 h suggested by Gerasimova. For the calcula- One-way analysis of similarity, ANOSIM
tion of daily ration some authors (Pakhomov et (Clarke and Warwick, 1994), was employed to test
al., 1996) have recommended the substitution of 24 the hypothesis of no differences in mesopelagic fish
 10 in Eq. (3) for species that have an active assemblage above and below 450 m. This subrou-
feeding period of 10 h. Other authors argue that tine compares the average rank similarities within
this substitution introduces a significant conserva- the predefined groups of samples with the average
tive bias (Williams et al., 2001). We calculated two similarity between groups. Values close to 1 indicate
alternative daily rations using both 10 and 24. a strong separation between groups, while a value
of 0 indicates no differences between groups.
2.3. Data analysis The similarity percentage routine (SIMPER)
(Clarke and Warwick, 1994), which was applied to
Density and biomass data for mesopelagic fish square-root transformed mesopelagic fish abun-
were calculated as individuals per filtered water dances, identified the contribution from individual
volume. The filtered volume was calculated by species to the dissimilarities between (the deep and
multiplying the trawled distance of the vessel with the shallow) sample groups.
the estimated mouth opening (200 m2) of the PT- Various univariate indices were calculated in
1088. order to characterize the species assemblages of the
Community structure was investigated with the deep and shallow group of samples: species
number, Shannon’s diversity index (H0 ) (Shannon
Primer-E5 Software package (Clarke and War-
wick, 2001). To reduce the weighting of dominant and Weaver, 1949) and Pielou’s Evenness Index (J)
species, the densities were square-root transformed (Pielou, 1975). As these indices are known to be
prior to the computation of the triangular influenced by sample size, we also calculated the
similarity matrices based on Bray-Curtis simila- taxonomic diversity D and taxonomic distinctness
DÃ ; which consider the taxonomic relatedness of
rities (Field et al., 1982). The results of the latter
were classified by hierarchical agglomerative clus- species (Warwick and Clarke, 1995). Taxonomic
diversity D is empirically related to H 0 but
ter analysis using the group average linking
method, and ordinated by a non-metric, multi- contains, in addition, information on the taxo-
dimensional scaling technique (MDS). nomic separation of the species in a sample, i.e.
The BIOENV sub-routine was used to relate the besides the distribution of individuals among
community patterns of the mesopelagic fish species it also takes into account the distribution
assemblage to six environmental variables: light of species in the taxonomic system by weighting
intensity (W/m2) (indicating time of day), mini- the co-occurrences of species according to the
mum and maximum values of sample depth, degree of separation in the hierarchical classifica-
bottom depth, temperature and salinity at the tion (1=different species, 2=different genera,
3=different families, 4=different orders). DÃ is
sampled depth horizon. Values of average light
ARTICLE IN PRESS
C. Pusch et al. / Deep-Sea Research I 51 (2004) 1685–1708
1690
derived from D but measures solely the taxonomic Sea and Scotia Sea mingle. Analysis of vertical
distinctness of species in a sample, without the temperature and salinity sections through the
contribution from species diversity. study site revealed the existence of two different
Oceanographic data were obtained from the water masses in the slope region of King
program Ocean Data View (Schlitzer, 2003). The George Island (Fig. 2A–B). The water mass in
results are presented as a section through the study the upper 200 m was composed of ASW, char-
area. acterized by a temperature minimum in 50–150 m
depth (To0:5 1CÞ and a low salinity (o 34.4). The
origin of this water body is due to the cooling of
the surface water during winter. The properties of
3. Results
this so-called winter water are stable year round
(Hofmann et al., 1996). During our study the
3.1. Oceanographic conditions
surface water temperature was slightly
increased by enhanced solar radiation. The
The study area is situated in the Weddell-Scotia-
dominating water mass below 400 m depth,
Confluence, where water masses of the Weddell
Fig. 2. (A–B) RV ‘‘Polarstern’’ cruise ANT XIV/2, east–west hydrographic section through the study area as indicated in Fig. 1, (A)
temperature (1C) and (B) salinity.
ARTICLE IN PRESS
C. Pusch et al. / Deep-Sea Research I 51 (2004) 1685–1708 1691
Circumpolar Deep Water, was characterized by other families represented by just one species.
Three myctophids, Gymnoscopelus fraseri, G.
higher temperature (T40.5 1C) and salinity
hintonoides and Protomyctophum choriodon, were
(S434:5). A transition zone between these water
masses occupies the depth range 200–400 m, in newly recorded for the South Shetland Island
which the gradients of temperature and salinity region.
were strongest. Myctophids dominated the mesopelagic fish
community by number and wet weight accounting
for 98.5% of all sampled individuals and 87.3% of
3.2. Species composition
the total weight. E. antarctica was the most
abundant species by number and weight. This
The total catch of mesopelagic fish taken on 15
myctophid accounted for 60.7% of all sampled
pelagic trawl stations included 16 343 specimens
individuals, followed by P. bolini (19.6%), G.
with a wet weight of 122.5 kg (Table 2). Eighteen
braueri (13.8%) and G. nicholsi (4.0%). All other
mesopelagic species belonging to 10 families and
mesopelagic fish species were of numerically minor
13 genera were identified. The family Myctophidae
importance (o2% in total).
was by far the most speciose (9 species), with all
Table 2
Number of individual fish, density as number of individuals per 105 m3 filtered volume, wet weight in kg and biomass as g wet weight
per filtered volume in 15 PT-1088 samples
Species Abundance Biomass
Ind/105 m3 g ww/105 m3
No. Inds. (%) kg (%)
Astroneshtidae
Borostomias antarcticus 1 (0.01) 0.00 0.01 (0.01) 0.05
Bathylagidae
Bathylagus antarcticus 4 (0.02) 0.03 0.05 (0.04) 0.43
Chiasmodontidae
Chiasmodon niger 1 (0.01) 0.00 0.02 (0.02) 0.09
Gempylidae
Paradiplospinus gracilis 182 (1.11) 1.14 14.96 (12.21) 93.29
Gonostomatidae
Cyclothone sp. 2 (0.01) 0.02 0.01 (0.01) 0.09
Melamphaidae
Poromitra crassiceps 2 (0.01) 0.01 0.02 (0.02) 0.09
Microstomatidae
Nansenia antarctica 3 (0.02) 0.02 0.09 (0.07) 0.66
Myctophidae
Electrona antarctica 9931 (60.71) 59.48 67.93 (55.46) 407.98
Gymnoscopelus braueri 2253 (13.77) 20.23 13.71 (11.19) 122.72
Gymnoscopelus fraseri 2 (0.01) 0.01 0.02 (0.02) 0.14
Gymnoscopelus hintonoides 15 (0.09) 0.11 0.13 (0.11) 0.99
Gymnoscopelus nicholsi 647 (3.96) 4.2 21.15 (17.27) 136.81
Gymnoscopelus opisthopterus 27 (0.17) 0.13 0.96 (0.78) 4.52
Krefftichthys anderssoni 23 (0.14) 0.18 0.07 (0.06) 0.55
Protomyctophum bolini 3212 (19.63) 20.1 2.97 (2.42) 18.65
Protomyctophum choriodon 1 (0.01) 0.00 0.01 (0.01) 0.04
Notosudidae
Scopelosaurus hamiltoni 3 (0.02) 0.02 0.27 (0.22) 1.48
Paralepididae
Notolepis coatsi 34 (0.21) 0.22 0.11 (0.09) 0.71
Total 16 343 122.5
ARTICLE IN PRESS
C. Pusch et al. / Deep-Sea Research I 51 (2004) 1685–1708
1692
In wet weight, E. antarctica accounted for made. Nevertheless, some assumptions about the
55.6% of total mesopelagic fish biomass, far vertical distribution patterns of the more abundant
exceeding G. nicholsi (17.3%), G. braueri (11.2) species are possible. All four myctophids were
and the gempilyd, Paradiplosinus gracilis (12.2%). more abundant, in the 273–825 m depth range, in
daytime samples than in nighttime samples (Fig.
3A–D). This observation supports the results of
3.3. Patterns of vertical distribution
earlier studies that the bulk of Antarctic mycto-
phids migrate to epipelagic layers at night (Torres
Because of the sampling strategy (no samples
and Somero, 1988; Lancraft et al., 1989).
above 273 m depth and only four night tows) no
P. gracilis was concentrated in depths below
detailed conclusions about the vertical migration
400 m during daytime (Fig. 3E). The distribution
behaviour of the mesopelagic community can be
(B) G. braueri
0
0 (A) E. antarctica
200
200
400
400
600
600
800
800
300 200 100 0 100 200 300
300 200 100 0 100 200 300
(D)
0 0
(C)
G. nicholsi P. bolini
200 200
Depth (m)
400 400
600 600
800 800
20 10 10 20 80 40 0 40 80
0
(E)
0
P. gracilis
200
400
600
800
4
4 2 0 2
Density
Fig. 3. (A–E) Vertical day/night distribution: (A) E. antarctica, (B) G. braueri, (C) G. nicholsi, (D) P. bolini, (E) P. gracilis, Density as
individuals per 105 m3; open bars=day tows; filled bars=night tows.
ARTICLE IN PRESS
C. Pusch et al. / Deep-Sea Research I 51 (2004) 1685–1708 1693
data show an increased density in this species
between 400 and 500 m, suggesting vertical migra-
tion behaviour at least in mesopelagic depths. The
four myctophids showed different daytime depth
distributions: The bulk of E. antarctica individuals
were found below 400 m depth (Fig. 3A) and the
minimum depth of occurrence of G. braueri was
750 m (Fig. 3B). In contrast, P. bolini occurred
over a depth range of 273–825 m, with the bulk of
the population in the shallow 273–450 m interval
(Fig. 3D). G. nicholsi occurred over the whole Fig. 4. Dendrogram representing the classification of 15 PT-
1088 stations according to mesopelagic fish assemblage; Station
sampled depth range (273–825 m) with a centre of
number and maximum sampled depth are indicated. Hatching
distribution between 400 and 597 m (Fig. 3C).
indicates shallow stations (depth 302–450 m).
3.4. Community analysis
Deep
Shallow
Cluster analysis based on densities of mesope-
lagic fish showed a clear separation of stations at 68
an arbitrary level of 40% (Fig. 4). The first cluster
65
comprises stations sampled at and above 450 m 70
depth. The second cluster is exclusively composed 58 54
60 59
69
of stations taken in depth4450 m, with the 52
53 55
57
exception of station 60, where sampling depth 63
was 380–440 m. 64
The separation of stations at 450 m depth is
confirmed by the ordination with MDS of the
51
same assemblage data (Fig. 5). Again, station 60
was more closely associated with the deeper group
of stations. Fig. 5. MDS plot for 15 PT-1088 stations according to the
mesopelagic fish assemblage; shallow and deep refer to the
An exploratory analysis, BIOENV, was applied
sampled depth; shallow=depth 302–450 m; deep=depth
to examine which abiotic variables could best
465–825 m; stress (=goodness of fit)=0.07.
explain the observed patterns in the mesopelagic
fish assemblage. The maximum matching coeffi-
cient was achieved by the abiotic variable max- An analysis of similarity (ANOSIM) was
imum sample depth (51.7%) and supported performed to test for statistical differences in
therefore the result from the MDS and cluster species composition between shallow stations
analyses. Minimum sample depth (46.4%) and (295–450 m depth) and deep stations
(465–825 m). A value of R ¼ 0:561 supported the
minimum and maximum salinity (34.9% and
28.6%, respectively) were further useful abiotic results of the classification and ordination of the
parameters to explain the observed pattern in the data and indicated significant differences in species
mesopelagic fish community. Other abiotic vari- composition between shallow and deep stations
ables showed a low correlation with the biotic (P ¼ 0:003).
similarity matrix and yielded matching coefficients The similarity percentage procedure (SIMPER)
below 25%. Sample depth is thus the best was applied to identify those species that con-
environmental variable to explain the grouping tribute most to the observed differences between
of the samples in a manner consistent with the shallow and deep samples (Table 3). Only three
biotic pattern. species were more abundant in the shallow group
ARTICLE IN PRESS
C. Pusch et al. / Deep-Sea Research I 51 (2004) 1685–1708
1694
Table 3
SIMPER analysis; discriminating species between ‘‘deeper’’ and ‘‘shallower’’ station groups as revealed by cluster analysis
Avg. density (Ind/105 m3)
Species Cumulative
Shallow Deep Contribution (%) Contribution (%)
Electrona antarctica 19.01 94.89 43.33 43.33
Protomyctophum bolini 33.6 8.3 18.8 62.13
Gymnoscopelus braueri 0.01 37.93 16.35 78.48
Gymnoscopelus nicholsi 4.13 4.26 5.89 84.36
Paradiplospinus gracilis 0.3 1.88 5.22 89.59
Notolepis coatsi 0.05 0.37 2.29 91.88
Krefftichthys anderssoni 0 0.34 2.28 94.16
Gymnoscopelus hintonoides 0 0.2 1.67 95.82
Gymnoscopelus opisthopterus 0 0.24 1.19 97.02
Bathylagus antarcticus 0 0.06 0.5 97.52
Nansenia antarctica 0.01 0.03 0.47 97.99
Scopelosaurus hamiltoni 0 0.03 0.46 98.45
Gymnoscopelus fraseri 0 0.03 0.41 98.85
Protomyctophum choriodon 0 0.02 0.28 99.13
Poromitra crassiceps 0.01 0 0.25 99.38
Cyclothone sp. 0 0.04 0.25 99.62
Chiasmodon niger 0 0.01 0.2 99.82
Borostomias antarcticus 0.01 0 0.18 100
Species are ordered in decreasing contribution (%) to the total dissimilarity.
Table 4
of samples, while 15 species were more abundant
Univariate indices of the deeper and shallower groups of
in the deeper group. The myctophid P. bolini was
stations over the slope of King–George Island
the best indicator species for samples taken above
450 m depth and accounted for 18.8% of the Shallow Deep
observed differences. Deeper stations were char-
No. of species 9 16
acterized by relatively higher abundances of E.
Density (Ind./105 m3) 57.13 148.61
antarctica and G. braueri and P. gracilis , which Pielou’s evenness (J) 0.35 0.41
together accounted for 65% of the observed Shannon’s diversity (H0 ) 0.51 0.79
differences between the two groups of samples. Taxonomic diversity (D) 10.03 14.64
Taxonomic distinctness (DÃ ) 35.87 37.02
The calculation of several univariate indices for
the species densities above and below 450 m depth
reflected the observed differences between the
not significant. The taxonomic distinctness DÃ
mesopelagic fish assemblages (Table 4). The
number of species (9) in depth between 273 and showed a high similarity between shallow and
450 m was lower, compared to 16 species in depths deep samples.
below 450 m (two individuals of Cyclothone sp.
3.5. Feeding ecology
were not identified to species level). Also the
average density of mesopelagic fish was 2.5 times
3.5.1. Food composition
higher in the deeper group samples. Shannon’s
diversity (H0 ), Pielou’s evenness index (J) and the The diet of E. antarctica was dominated by
taxonomic diversity D were higher in samples euphausiids in our samples. By number, euphau-
taken below 450 m depth. However, the observed siids (47.8%) are followed by ostracods (15.6%)
differences in diversity indices between the deeper and then by copepods (26.7%) (Fig. 6A; Table 5).
and shallower group of samples were statistically By dry weight (DW), euphausiids account for
ARTICLE IN PRESS
C. Pusch et al. / Deep-Sea Research I 51 (2004) 1685–1708 1695
G. braueri (n = 21)
E. antarctica (n = 61)
DW (%) A (%)
DW (%) N (%)
100 80 60 40 20 0 20 40 60 80 100
100 80 60 40 20 0 20 40 60 80 100
20 20
Eu (RI = 61.1) Eu (RI = 52.9)
40 40
60 60
Cop (14.9)
80 80 Os (19.0)
Os (12.0)
100 100
Co (13.9)
Am (6.0)
Ch (4.3)
Am (9.3)
Ch (4.9)
F (%) (B)
(A) F (%)
G. nicholsi (n = 135) P. bolini (n = 101)
DW (%) N (%) DW (%) N (%)
100 80 60 40 20 0 20 40 60 80 100 100 80 60 40 20 0 20 40 60 80 100
20 20
Eu (RI = 44.0)
40 40 Co (RI = 82.0)
60 60
80 80
Co (37.9)
100 100 Os (10.0)
Eu (6.2)
Os (7.3)
F (%)
(D)
(C)
Ch (5.7)
Pi (4.7)
F (%)
P. gracilis (n = 28)
DW (%) N (%)
100 80 60 40 20 0 20 40 60 80 100
Am: Amphipods
Ceph: Cephalopds
20
Ch: Chaetognaths
Pi (RI = 76.8) 40
Co : Copepods
60
Eu : Euphausiids
80
Os : Ostracods
Eu (15.1)
100
Pi : Pisces
Ceph (8.4)
(E) F (%)
Fig. 6. (A–E) Percentage composition of the main prey items of (A) E. antarctica (B) G. braueri (C) G. nicholsi (D) P. bolini (E) P.
gracilis, by percent dry weight (DW), percentage number (N) and frequency of occurrence (F), The relative importance (RI) Index is
presented by the size of the respective rectangles. n=number of stomachs containing food.
1696
Table 5
Diet composition of E. antarctica, G. braueri, G. nicholsi, P. bolini and P. gracilis, showing number and reconstructed dry weight of each food item; the respective
percentage of the total number and the total dry weight is given in parenthesis; ‘‘-’’: absent.
Electrona antarctica Gymnoscopelus braueri Gymnoscopelus nicholsi Protomyctophum bolini Paradiplospinus gracilis
SL range (mm) 57–113 69–121 123–172 33–53 284–469
No. of fish examined 145 67 186 122 100
No. of stomach empty (%) 84 (57.9) 46 (68.7) 51 (27.4) 21 (17,2) 72 (72)
C. Pusch et al. / Deep-Sea Research I 51 (2004) 1685–1708
No. (%) Weight (%) No (%) Weight (%) No (%) Weight (%) No. (%) Weight (%) No. (%) Weight %
Amphipoda 2 (2.2) 210.86 (13.9) 1 (3.2) 84.08 (23.5) 1 (0.1) 50.75 (0.8) — — — — — — — —
Cephalopoda — — — — — — — — — — — — — — 3 (8.3) 9900 (7.4)
Psychroteuthis glacialis 3 (8.3) 9900 (7.4)
ARTICLE IN PRESS
Chaetognatha 5 (5.6) 3.9 (0.3) 2 (6.5) 1.56 (0.4) 21 (2.6) 241.78 (3.8) 4 (0.3) 3.12 (0.6) — — — —
Copepoda 24 (26.7) 5.42 (0.4) 7 (22.6) 1.97 (0.6) 517 (63.8) 235.67 (3.7) 1515 (96.6) 354.93 (84.2) — — — —
Calanus propinquus 1 (1.1) 0.8 (0.1) — — — — 23 (2.8) 21.84 (0.3) 8 (0.5) 6.21 (1.5) — — — —
Gaidius spp. 1 (1.1) 0.37 (0.0) 3 (9.7) 1.11 (0.3) 10 (1.2) 3.7 (0.1) 11 (0.7) 4.07 (1.0) — — — —
Metridia gerlachei 11 (12.2) 1.62 (0.1) — — — — 341 (42.1) 68.58 (1.1) 1438 (91.7) 319.83 (75.9) — — — —
Pareuchaeta spp. 1 (1.1) 1.35 (0.1) — — — — 67 (8.3) 95.82 (1.5) 13 (0.8) 15.87 (3.8) — — — —
Rhincalanus gigas — — — — 1 (3.2) 0.43 (0.1) 46 (5.7) 29.69 (0.5) 5 (0.3) 2.49 (0.6) — — — —
Unidentified calanoids 10 (11.1) 1.28 (0.1) 3 (9.7) 0.43 (0.1) 30 (3.7) 16.04 (0.3) 40 (2.6) 6.46 (1.5) — — — —
Euphausiacea 43 (47.8) 1292.07 (85.0) 13 (41.9) 268.66 (75.0) 226 (27.9) 4950.82 (77.9) 8 (0.5) 54.79 (13.0) 8 (22.2) 280.69 (0.2)
Euphausia cristallorophias — — — — — — — — 1 (0.1) 7.55 (0.1) — — — — — — — —
E. frigida 1 (1.1) 14.63 (1.0) — — — — 2 (0.2) 78.12 (1.2) — — — — — — — —
E. superba 14 (15.6) 807.23 (53.1) — — — — 67 (8.3) 3707.89 (58.3) — — — — 6 (16.7) 223.19 (0.2)
E. triacantha 1 (1.1) 19.96 (1.3) 1 (3.2) 36.82 (10.3) 4 (0.5) 163.74 (2.6) — — — — 1 (2.7) 45.4 (0.0)
Euphausia spp. 22 (24.4) 425.04 (28.0) 12 (38.7) 231.84 (64.8) 20 (2.5) 386.4 (6.1) — — — — — — — —
Thysanoessa macrura 5 (5.6) 25.21 (1.7) — — — — 132 (16.3) 607.12 (9.6) 8 (0.5) 54.79 (13.0) 1 (2.7) 12.1 (0.0)
Gastropoda 1 (1.1) 1.56 (0.1) — — — — — — — — 1 (0.1) 1.56 (0.4) — — — —
Ostracoda 14 (15.6) 4.78 (0.3) 8 (25.8) 1.74 (0.5) 39 (4.8) 8.76 (0.1) 41 (2.6) 7.13 (1.7) — — — —
Pisces — — — — — — — — 6 (0.7) 866.97 (13.6) — — — — 25 (69.4) 123868.28 (92.4)
Electrona antarctica — — — — — — — — — — — — — — — — 2 (5.6) 6435.36 (4.8)
Gymnoscopelus nicholsi — — — — — — — — — — — — — — — — 1 (2.7) 10162.92 (7.6)
Gymnoscopelus braueri — — — — — — — — — — — — — — — — 1 (2.7) 2270 (1.7)
Neopageotopsis ionah — — — — — — — — 2 (0.3) 408.25 (6.4) — — — — — — — —
Protomyctophum bolini — — — — — — — — 4 (0.5) 458.72 (7.2) — — — — — — — —
Polychaeta 1 (1.1) 1.84 (0.1) — — — — — — — — — — — — —
Total 90 1520.43 31 358.01 810 1742.7 1569 7.13 36 134048.97
ARTICLE IN PRESS
C. Pusch et al. / Deep-Sea Research I 51 (2004) 1685–1708 1697
85.0% of the total stomach content, far exceeding the identifiable food items respectively, while other
amphipods (13.9%) and the other prey categories food items accounted for less than 5% each. In dry
(o 2% in total). Euphausiids occurred in 65.6% of weight (DW) euphausiids made up 77.9% of the
all filled stomachs, while ostracods and copepods total prey, followed by fish (13.6%), chaetognaths
occurred less frequently (21.1% and 23.0%, (3.8%) and copepods (3.7%). The frequencies of
respectively). The RI values were 61.1% for occurrence for copepods, euphausiids, ostracods
euphausiids, 14.9% for copepods and 12.0% for and chaetognaths were 77.7%, 62.9%, 22.9% and
ostracods, stressing the importance of euphausiids 15.5%, respectively, while the remaining prey
in the diet of E. antarctica. categories accounted for less than 4.4% in total.
The size distribution of the 145 individuals The RI values were 44.0% for euphausiids and
examined for food composition was unimodal and 37.9% for copepods, while the remaining prey
ranged from 57 to 113 mm (Fig. 7A). To investi- categories accounted foro10%.
gate an ontogenetical shift in the prey size of E. The size distribution of the 186 individuals of G.
antarctica we compared the ratio of euphausiids nicholsi examined for stomach content was uni-
(largest prey item) to the total dry weight of modal and ranged from 123 to 172 mm (Fig. 7C).
stomach content in individuals smaller and larger The ratio of euphausiids (largest prey item) to the
than 80 mm SL. The ratio of euphausiids to the total dry weight of the stomach content was
54.7764.0% (avg. 7SD) for 123–144 mm SL
total dry weight of stomach contents was
60.7749.0% (avg.7SD) for 57–79 mm sized individuals and 58.3745.1% for 145–172 mm SL
individuals, and 68.6746.2% for 80–113 mm SL fish. The difference was not statistically significant
(Mann–Whitney U test, p40:05) and we therefore
fish. The difference was not significant (Man-
n–Whitney U-test, p40:05). Therefore, no signifi- suggest that no substantial ontogenetic shift in
cant shift in prey size was detectable over the size prey size occurs over the size range examined.
Prey species composition showed that G. nicholsi
range examined.
In this study, feeding intensity of E. antarctica fed on all five euphausiids species occurring in the
was low, with 57.9% of stomachs empty (Table 5). Southern Ocean (Table 5). The most important
euphausiid by food weight was E. superba (58.3%
More then half of all euphausiids were not
of total food weight), while Metridia gerlachei
identifiable to species. Nevertheless, krill (E.
superba) was the most important euphausiid in dominated numerically (42.1%). G. nicholsi was
the diet of this myctophid, accounting for more the only myctophid in our study that preyed on
other mesopelagic fish (Neopageotopsis ionah and
than one-half (53.1%) of the overall food weight.
The prey composition of G. braueri was Protomyctophum bolini).
The food spectrum of P. bolini, a relatively
similarly dominated by euphausiids (Fig. 6B,
Table 5). Euphausiids were most important by small-sized myctophid (SL 33–53 mm, Fig. 7D),
number (N ¼ 41:9%), DW (75.0% of total weight) was dominated by copepods (Fig. 6D, Table 5).
The N, DW and F values for copepods were
and frequency of occurrence (61.9%). The RI
reflected the high dietary value of euphausiids 96.6%, 84.2% and 91.1%, respectively. The RI
(52.9%), while ostracods, copepods and amphi- value of copepods was 82.0% followed by
pods were of minor importance with RI values of ostracods (10.0%) and euphausiids (6.2%).
19.0%, 13.9% and 9.3%, respectively. Despite the The prey species composition showed that
copepods of the genus M. gerlachei were by far
high significance of euphausiids in the diet, the
most abundant euphausiid, E. superba, was lack- the most important prey item, accounting for
ing (Table 5). The feeding intensity of G. braueri 91.7% of all identified food items (Table 5).
The diet of P. gracilis, the only non-myctophid
was noticeably low, with 68.7% of stomachs
containing no identifiable food. in our analysis, consisted almost exclusively of fish
The diet of G. nicholsi, was numerically domi- (Fig. 6E, Table 5). The three trophic parameters N,
DW and F for this prey category were 69.4%,
nated by copepods and euphausiids (Fig. 6C,
Table 5), accounting for 63.8% and 27.9% (N) of 92.4% and 82.1%, resulting in an RI of 76.8%.
ARTICLE IN PRESS
C. Pusch et al. / Deep-Sea Research I 51 (2004) 1685–1708
1698
(B)
(A)
100 100
E. antarctica G. braueri
n = 582 n = 187
80 80
60 60
40 40
20 20
0 0
0 20 40 60 80 100 120140 160180 200 0 20 40 60 80 100120140 160180 200
(D)
100
100 (C)
G.nicholsi P. bolini
n = 299 n = 189
80
80
Number of Fish
60
60
40
40
20
20
0
0
0 20 40 60 80 100120140 160180 200
0 20 40 60 80 100120 140160180 200
Length (SL, mm)
100 (E)
P. gracilis
n = 105
80
60
40
20
0
0 20 40 60 80 100120 140160180 200
Length (SL, cm)
Fig. 7. (A–E) Length-frequency plots of (A) E. antarctica (B) G. braueri (C) G. nicholsi (D) P. bolini (E) P. gracilis showing size classes
of fish measured (open bars) and sampled for diet analysis (filled bars); n total number of fish measured.
ARTICLE IN PRESS
C. Pusch et al. / Deep-Sea Research I 51 (2004) 1685–1708 1699
The average and range of SCI for G. nicholsi
Most fishes found in the stomachs were not
identifiable because of the advanced stage of were calculated to be 0.29% and 0.12% to 0.55%
digestion. Nevertheless, the common myctophids (Fig. 9A). There was no statistical difference in
E. antarctica, G. nicholsi and G. braueri could be feeding intensity between samples taken at day and
U-test, p ¼ 0:178).
identified. Additional euphausiids and cephalo- night (Mann–Whitney
pods (Psychroteuthis glacialis) were found in the However, feeding intensity was increased during
stomachs of P. gracilis, accounting for RI values evening times at 2100–2200 h (average SCI 0.38%)
of 15.1% and 8.4%, respectively. and reached a maximum at 2300–2400 h (SCI
0.55%). The observation of intensified feeding at
3.5.2. Feeding chronology these time periods is supported by a higher
The influence of sample depth on feeding proportion of undigested food items. We observed
intensity was investigated by the calculation of a decrease in feeding intensity after midnight and
the Stomach Content Index (SCI) for E. antarc- lasting until the early morning hours. The sto-
tica, G. nicholsi and P. bolini. We compared the machs of G. nicholsi contained fresh food items
SCIs of individuals taken in shallow (sample depth during day and night (stage 1) (Fig. 9C). For this
295–440) and deep tows (465–825 m) and found no reason we suggest that feeding activity was taking
differences for E. antarctica, G. nicholsi and P. place continuously.
bolini (Mann–Whitney U-Test; p40:05). There- In the small P. bolini average SCI (1.1%) was
higher than in G. nicholsi (Fig. 9B). We excluded
fore, it was concluded that sample depth had no
major influence on the SCI of these species. G. one station (58) with extraordinarily high
braueri was taken in too low numbers in the upper stomach filling (SCI 10.6%), and which we
depth horizon to be able to draw conclusions interpret as an outlier. No difference in median
about the influence of sample depth on feeding SCI was found between daytime and nighttime
samples (Mann–Whitney U test p ¼ 0:127). Never-
behaviour.
The SCI of E. antarctica fluctuated between theless, feeding activity was increased during hours
0.0% and 0.9% (avg. 0.47) over a 24 h time period of darkness when SCI values were above the
(Fig. 8A). Highest feeding activity was observed at average. A second peak in feeding intensity was
night (0200–0300 h). Minimum values in feeding observed in the morning hours from 0800 to
intensity were found for the intervals 0800–0900 h 0900 h (SCI 1.7%). During the day, SCI reached a
and 1100–1200 h. Nevertheless, no difference in minimum value of 0.21 at 1400–1500 h. The
feeding intensity was found between day and night highest proportions of undigested food items
samples (Mann–Whitney U-test, p ¼ 0:081), be- were recorded in the evening and before midnight
cause of the high variability of the SCI values (Fig. 9D).
during day. The proportion of undigested food
3.5.3. Daily ration and the impact of predation on
items (Stage 1) was reduced during daytime except
the krill stock
at one station during the time interval 1100–1200 h
where results are based on a low number of fish (5) The daily ration estimate was highest for the
small myctophid P. bolini, accounting for 2.48%
and food items (2) (Fig. 8C).
The average SCI of G. braueri was 0.22%, and of fish wet weight assuming a 10 h feeding period
values ranged from 0.04% to 0.4% (Fig. 8B). and 5.94% assuming that the species feeds 24 h a
It is difficult to draw conclusions about the day (Table 6). The lowest daily ration was
calculated for G. braueri, accounting for 0.5%
feeding chronology of this species because of the
low overall feeding intensity. We found no wet body weight under the conservative assump-
statistical differences in SCI between day and tion and 1.19% with a 24 h feeding period. The
night samples (Mann–Whitney U-test, p ¼ 0:867). estimated daily rations of G. nicholsi are calculated
to be only slightly higher than in G. braueri at
The proportion of fresh prey items was highest
0.65% and 1.57%. E. antarctica took an inter-
in the early evening (time interval 2000–2100 h)
(Fig. 8D). mediate position with daily rations of 1.06% and
ARTICLE IN PRESS
C. Pusch et al. / Deep-Sea Research I 51 (2004) 1685–1708
1700
Fig. 8. (A–D) Diel changes in stomach content of E. antarctica and G. braueri: (A–B) Average Stomach Content Index (SCI); vertical
bars=standard error; n=number of individuals examined. (C–D) Proportion of digestion stage (stage 1=fresh food, stage 2=slightly
digested, stage 3=only indigestible remains). Sunset (SS) and sunrise (SR) are indicted on the x-axis. Filled bars represent hours of
darkness.
2.54% of wet body weight by the 10 and 24 h region in the years 1996/1997 was estimated to be
92 g/1000 m3 (Siegel et al., 1998b). We estimated
feeding period assumption.
Two myctophids in our study, E. antarctica and the biomass of E. antarctica and G. nicholsi to be
4.08 and 1.36 g/1000 m3, respectively, in slope
G. nicholsi, fed on krill (E. superba). Based on the
food composition data and the daily rations, we waters off King George Island. The daily intake
of krill by E. antarctica was assessed to be 1.06%
assessed the predation impact of these myctophids
on the krill stock in the study area. of its body wet weight during a 10 h feeding period
E. superba was the most important prey item in and 2.54% in 24 h.
the diet of E. antarctica and G. nicholsi and The predation impact of G. nicholsi on krill was
accounted for 53.1% and 58.3% of the total food estimated to be 0.65% of body wet weight
weight in these species, respectively. The standing under the assumption of a 10 h feeding period
stock of krill biomass in the Elephant Island and 1.57% assuming a 24 h feeding period.
ARTICLE IN PRESS
C. Pusch et al. / Deep-Sea Research I 51 (2004) 1685–1708 1701
Fig. 9. (A–D) Feeding periodicity of G. nicholsi and P. bolini, same abbreviations as in Fig. 8.
Neglecting seasonal variability in consumption, micronekton community in the Ice-free Zone of
our data suggest an annual removal of 9.0–21.8% the Scotia Sea (Lancraft et al., 1989; Piatkowski
of the krill stock by E. antarctica and 2.1–5.0% by et al., 1994).
G. nicholsi. It was expected that the mesopelagic fish
assemblage at King George Island, which is
localized in the Seasonal Pack-ice Zone, would
be characterized by a lower species number and
4. Discussion
diversity compared to the Ice-free Zone (Kock,
4.1. Species composition 1992). The relatively high number of species over
the slope of King George Island could be due to
Eighteen mesopelagic fish species were identified increased productivity over the slope of King-
in samples in the slope waters of King George George Island, itself due to a shelf-break front as
Island, situated in the seasonal pack ice zone. This described by Ichii et al. (1998). Unpublished
number agrees with previous studies of the LIDAR measurements performed on cruise ANT
ARTICLE IN PRESS
C. Pusch et al. / Deep-Sea Research I 51 (2004) 1685–1708
1702
Table 6
Daily ration and predation impact on the krill stock of four myctophid species
E. antarctica G. braueri G. nicholsi P. bolini
Avg. SCI (% body dry weight) 0.47 0.22 0.29 1.10
Feeding period (h) 10/24 10/24 10/24 10/24
Daily ration (% body dry weight) 0.59/1.41 0.28/0.66 0.36/0.87 1.38/3.3
Daily ration (% body wet weight)a 1.06/2.54 0.50/1.19 0.65/1.57 2.48/5.94
Proportion of E. superba in food weight (%) 53.1 0.00 58.30 0.00
Daily ration of E. superba (% body wet weight) 0.56/1.35 0.00 0.38/0.92 0.00
Predator density (g/1000 m3) 4.08 1.20 1.36 0.19
Daily consumption of krill (mg/1000 m3) 22.85/55.08 0.00 5.17/12.51 0.00
Annual consumption of krill (g/1000 m3) 8.32/20.05 0.00 1.88/4.55 0.00
Percentage of krill stock removed (%)b 9.04/21.79 0 2.05/4.95 0
Daily rations were calculated for both 10 and 24 h periods of feeding.
a
Conversion factor daily ration dw to ww 1.8.
b
Taking in account a krill biomass of 92 g/1000 m3 (Siegel et al., 1998b).
XIV/2 indicated an increased chlorophyll-a con- Front (Rowedder, 1979b; Hulley, 1981; McGinnis,
centration in the vicinity of the Scotia-Weddell 1982; Lancraft et al., 1989; Hoddell et al., 2000).
The size spectrum of E. antarctica suggested the
Confluence, which passed through the study site.
A higher diversity of mesopelagic fish, resulting presence of the age classes two and three during
from an ‘island effect’, was described by Piat- our study, according to the results of Greely et al.
kowski et al. (1994) for an assemblage above the (1999).
The centre of distribution of P. choriodon, G.
slope of South Georgia when compared to an
fraseri and G. hintonoides are known to be in the
oceanic station in the northern Scotia Sea. The
snake mackerel P. gracilis is a characteristic area of the South Polar Front and north of it
species of slope habitats with benthopelagic (Hulley, 1981). The first records of these species in
life style (Nakamura, 1990) which was also the South Shetland Islands region give support to
recorded over the slope of South Georgia the idea that frontal zones, such as the South Polar
(Piatkowski et al., 1994). Front Zone, are not impervious barriers but may
However, our study supports the results of be penetrated by mesopelagic fish that are trapped
previous studies that characterize the Antarctic in eddies (Kock, 1992).
micronekton community as a low diversity oceanic
faunal assemblage, compared to subtropical and
4.2. Community structure
tropical oceans, where species numbers are four to
five times higher (Hopkins and Lancraft, 1984;
Lancraft et al., 1989). Our data showed a strong separation between
The dominant position of E. antarctica, G. samples taken above and below 450 m depth.
braueri, P. bolini in the Antarctic micronekton These finding could be well correlated with the
assemblage have also been shown for the Scotia presence of warmer and denser CDW at and below
Sea and slope waters off South Georgia (Piat- this depth. Samples from the CDW showed a
kowski et al., 1994). These species are assumed to higher species diversity based on Shannon’s
be the most common myctophids of the Southern diversity and Taxonomic diversity (D). In contrast
the Taxonomic distinctness (DÃ ) showed a high
Ocean (Hulley, 1981). The high densities of E.
antarctica over the slope of King George Island degree of similarity between the deeper and lower
confirmed the importance of this species in the mesopelagic fish community, due to the dom-
pelagic ecosystem south of the Antartic Polar inance of the family Myctophidae and the high
ARTICLE IN PRESS
C. Pusch et al. / Deep-Sea Research I 51 (2004) 1685–1708 1703
number of congeneric species in the lower meso- et al., 1998). In our study, no influence of the shelf-
pelagic assemblage. break on the mesopelagic fish assemblage was
The higher density of mesopelagic fish samples discernable. This could be due to the relatively
taken in the CDW compared to ASW are in small horizontal extent of our study site. There-
accordance with the results of Piatkowski et al. fore, further studies of community structure
(1994), who found a high similarity in species should cover a wider geographical area to fully
composition in lower mesopelagic depths encompass hydrographic features like frontal
(400–1000 m) sampled at two different stations in zones.
the Scotia Sea, while samples taken in the upper
4.3. Feeding ecology
depth horizons were clearly separated geographi-
cally. They found the highest biomass and
4.3.1. Food composition
abundance in the upper layers of the CDW
(400–800 m). Our data showed that Antarctic myctophids,
The analysis of the community structure of like myctophids in other oceans, are primarily
migratory species, like mesopelagic fish, has to zooplanktivores. They feed mainly on copepods
consider the day/night distribution of the species. and euphausiids. The proportions of these main
Because of our restricted sampling program (no prey categories differed in the diets of the species
examined. Large myctophids (460 mm SL), E.
samples above 273 m depth, only four night
antarctica, G. braueri and G. nicholsi fed mainly on
samples) conclusions about the vertical migration
euphausiids, while the smaller-sized P. bolini
patterns of the mesopelagic fish can be drawn only
from data in the literature (Lancraft et al., 1989; (53 mm maximum SL) preyed mainly on cope-
Piatkowski et al., 1994; Duhamel et al., 2000). pods. These results suggest niche separation by the
The results of the multivariate statistical analy- utilization of different prey taxa.
sis of the mesopelagic fish data can be best More than 90% of the krill biomass is concen-
explained by diel migratory behaviour of the trated in the upper 100 m of the water column in
species. The most characteristic species, E. antarc- the South Shetland Island region (Siegel, 1985).
tica and G. braueri, in the group of deeper stations The diet analysis indicated that krill was the main
prey item of E. antarctica and G. nicholsi. It can be
(as defined by SIMPER analysis) perform exten-
assumed that E. antarctica and G. nicholsi were
sive vertical migrations (Torres and Somero, 1988;
Lancraft et al., 1989; Piatkowski et al., 1994). sampled below their main feeding depth, because
During daylight, they stay in the core of the CDW the fishes examined in our study were taken in
(400–800 m) and migrate to the upper 200 m at depths greater than 273 m. This assumption was
supported by the low stomach filling in E.
night. A third species characteristic of deeper
stations was P. gracilis, a species whose vertical antarctica (57.9% empty stomachs) and the
distribution is probably confined to mesopelagic advanced digestion stage of krill found in the
stomachs of G. nicholsi (46.3% digestion stage 3).
depths. The most characteristic species of shallow
samples, P. bolini, also migrates vertically, but In contrast, the proportion of heavily digested
Thysanoessa macrura, a species characterized by a
over a more restricted range as described by
Piatkowski (1989) and Lancraft et al. (1989). wider vertical distribution range (Lancraft et al.,
Therefore, a substantial part of the population 1989), was much smaller (9.1% digestion stage 3).
The high importance of M. gerlachei in the diet of
remains in the ASW (0–400 m depth) during
P. bolini in our study re-confirmed the results of
daylight.
Duhamel et al. (2000) showed a clear change in Ascenio and Moreno (1984). The vertical distribu-
tion maximum of M. gerlachei is in the upper and
the species composition of mesopelagic fish in the
vicinity of the Polar Front at Kerguelen Island. A lower mesopelagic depths (Huntley and Escritor,
frontal system has also been described for the 1992). It can be assumed that this copepod species
shelf-break zone north of King George Island that was available over the whole vertical distribution
range of P. bolini. In contrast to these findings,
results in increased krill densities in this area (Ichii
ARTICLE IN PRESS
C. Pusch et al. / Deep-Sea Research I 51 (2004) 1685–1708
1704
Gaskett et al. (2001) described P. bolini as an E. antarctica taken in samples from November
to January (Rowedder, 1979a) and P. bolini
euphausiid feeder, as this prey category made up
more than the half of the food wet weight in their sampled in February (Ascenio and Moreno,
study that was undertaken close to the subantarc- 1984) have been shown to prey heavily on
tic Maquarie Island. Such differences support the development stages of krill. The breeding season
findings of Pakhomov et al. (1996), namely that of krill fluctuates in the Antarctic Peninsula region
the food composition of myctophids varies by from the beginning of December until the end of
locality and probably by season. The same March (Siegel et al., 1997). During our study, at
conclusion is true for the role of krill in the diets the beginning of December, most female krill
of E. antarctica and G. nicholsi. The importance of individuals were in a pre-spawning or early
adult krill in the diet of both species has been spawning condition (Siegel et al., 1998a). There-
described previously for several areas of the fore, it can be assumed that no krill development
Southern Ocean (Rembiszewski et al., 1978; stages were available in the zooplankton commu-
Rowedder, 1979a; Lubimova et al., 1983; Kozlov nity at the time of our study. In summary, our data
and Tarverdiyeva, 1989). However, Pakhomov et indicate that krill was an important prey item for
al. (1996) found no krill in the diet of E. antarctica E. antarctica and G. nicholsi over the slope of King
and only small quantities in the diet of G. nicholsi George Island. It can be expected that the dietary
sampled in the Prydz Bay, South Georgia and importance of krill for myctophids will be en-
Lazarev Sea (Indian Sector) regions. Therefore, hanced during the course of the austral summer,
these authors suggest that krill only plays a when krill abundance increases and their smaller
significant role in the diet of myctophids only development stages become available in the
during certain time periods and within specific zooplankton (Siegel et al., 1997; Ichii et al.,
1998). Our data also show that P. gracilis is an
areas. An additional reason for the minor im-
portance of krill in the diet of E. antarctica and G. important predator of mesopelagic fish over the
nicholsi in their study could be result of size- slope of King George Island, so confirming the
dependent feeding. Numerous studies have shown results of Rembiszewski et al. (1978).
that larger individuals of a population are able to
4.3.2. Feeding chronology
consume larger prey items, which is often demon-
Except for G. braueri, all myctophids in the
strated by an ontogenetic switch from copepods to
euphausiids as the main prey category (Young and present study showed diurnal feeding patterns
Blaber, 1986; Williams et al., 2001). The size- characterized by increased food intake during
ranges of E. antarctica (25–85 mm SL) and G. night. Nevertheless the day-night differences in
nicholsi (69–139 mm SL) examined by Pakhomov feeding activity were not significant for any
et al. (1996) were smaller than their size-ranges species. Pakhomov et al. (1996) suggested an
in our study, namely 57–113 mm SL and active nocturnal feeding period of 8–10 h for
123–172 mm, respectively. Nevertheless, the Antarctic myctophids. Migratory myctophids,
difference in diet composition cannot be explained especially in nutrient poor tropical and subtropical
entirely by predator length. In the present waters, are reported to show a pronounced
study the diet of the smaller-sized E. antarctica chronology of feeding activity, by feeding in the
(57–79 mm) and G. nicholsi (123–144 mm), that is food rich epipelagic layer at night (Clarke, 1978;
size ranges represented in Pakhomov’s samples, Kinzer and Schulz, 1985; Gartner et al., 1997). In
have a similar food composition to larger contrast, in more productive areas, a number of
migratory myctophids such as Diaphus theta,
individuals and is dominated by euphausiids.
Therefore scarcity of E. superba in the diet of Stenobrachius leucopsarus, Tarletonbeania crenu-
E. antarctica and G. nicholsi in Pakhomov’s laris (Tyler and Pearcy, 1975), Benthosema glaciale
(Kinzer, 1977), Diaphus danae, Lampanyctus hec-
study may be due more to the effect of food
toris (Young and Blaber, 1986) and D. theta
availability than the effect of size-dependent
feeding behaviour. (Moku et al., 2000) tend to feed continuously. It
ARTICLE IN PRESS
C. Pusch et al. / Deep-Sea Research I 51 (2004) 1685–1708 1705
should be noted that the Southern Ocean ecosys- examined were empty, Table 5). Because of this
tem is characterized by a seasonal high productiv- fact, our calculated daily rations could be under-
ity near the pack ice and in the shelf regions like estimated. However, we suggest that our estimates
the slope region of King George Island (Hempel, are realistic, because we have re-calculated prey
1985; Ichii et al., 1998). dry weights by length-weight regressions of un-
In our study, G. nicholsi showed an increased digested prey items. The use of re-calculated
feeding activity in the time period 2100–2400 h, weight is expected to bias the SCI values in a
while in P. bolini SCI was above the average from positive direction (Hyslop, 1980). It is for this
2000 to 0300 h. However, neither species ceased reason that we felt justified in applying the daily
feeding during the day, and both showed a second rations of myctophids in the calculation of their
feeding activity peak in the time interval predation impact on the krill stock in the study
0800–0900 h. A daily feeding period in P. bolini area. Most estimates of the predation impact by
appears to be possible due to the relatively shallow mesopelagic fish on the zooplankton stock have
daytime depth they occupy (273–500 m). This been performed in tropical and subtropical
range overlaps with the vertical distribution range waters. Despite differences in regions and
of its major prey item, M. gerlachei (Hernandez- methods of estimation used, all these studies
Leon et al., 2001). The SCI of E. antarctica indicate that mesopelagic fish are able to remove
increased during the hours of darkness and in two to three times the standing stock of herbivor-
the early morning. In contrast during summer and ous zooplankton (Dalpadado and Gjoesaeter,
autumn at South Georgia, South Sandwich and 1988; Hopkins and Gartner, 1992; Gartner et al.,
South Orkney Island Rowedder (1979a) observed 1997).
the highest feeding activity of this species in the There has been only one previous estimate of
afternoon (1400 h) and second peak in the morning zooplankton consumption by Antarctic mesopela-
(0600 h). The different feeding patterns of E. gic fish. Naumov (1985) as cited in Pakhomov et
antarctica observed by Rowedder suggest that al. (1996) on the basis of an assumed mesopelagic
the feeding chronology of mesopelagic fish can fish stock of 275 million tonnes in the Southern
show a high regional and seasonal variability. Ocean, estimated an annual consumption of 1085
million tonnes of meso- and macro zooplankton
4.3.3. Daily ration and predation impact on the krill (excluding E. superba). This equates to an esti-
stock mated predation impact equivalent to 40% of the
Our estimates of daily rations (0.7–3.3% of dry annual secondary production.
body weight, assuming a 24 h feeding period) were We have calculated the annual removal of the
krill stock in the Southern Ocean region by E.
in good agreement with the estimates of Pakho-
antarctica and G. nicholsi, which are thought to be
mov et al. (1996) for Antarctic myctophids. The
latter ranged from 0.2% to 4.4% of dry body two of the most abundant myctophids in the
weight. Southern Ocean, to be 11.1–26.7%. These results
The estimates of daily ration given by Rowedder confirm the important position of mesopelagic fish
(1979a) for E. antarctica (5% of dry body weight) in the Antarctic oceanic system, and give support
and by Gerasimova (1990) for E. carlsbergi to the suggestion of Lancraft et al. (1989) that they
(3.7–5.6 wet body weight) were higher than in are the most prevalent predators on krill in the
our results. While feeding activity in migratory Antarctic ecosystem. However, the figures given in
mesopelagic fish has been observed to reach the present study should be treated with caution
maximum values in the food-rich epipelagic layers since they represent a ‘‘snapshot’’ of the mid-water
(Gartner et al., 1997), the calculations of daily fish community over the slope of King George
rations in our study were based on individuals Island. Because the feeding impact of mesopelagic
sampled from mesopelagic depths. E. antarctica fish on the krill stock is largely dependent on
and G. braueri, in particular, showed a low feeding locality and season, future studies should be
activity (57.9% and 68.7% of all stomachs carried out at different seasons of the year to get
ARTICLE IN PRESS
C. Pusch et al. / Deep-Sea Research I 51 (2004) 1685–1708
1706
a more realistic picture of the Antarctic mesope- Eggers, D.M., 1977. Factors in interpreting data obtained by
diel sampling of fish stomachs. Journal of the Fisheries
lagic food web.
Research Board of Canada 34, 290–294.
Field, J.G., Clarke, K.R., Warwick, R.M., 1982. A practical
strategy for analysing multispecies distribution patterns.
Acknowledgements Marine Ecology-Progress Series 8, 37–52.
Gartner Jr., J.V., Crabtree, R.E., Sulak, K.J., 1997. Feeding at
Special thanks to Dr. Volker Siegel (Sea Fisheries depth. In: Randall, D.J., Farrel, A.P. (Eds.), Deep-Sea
Fishes. Academic Press, San Diego, pp. 1–388.
Research Institute) for his great assistance in the
Gartner Jr., J.V., Conley, W.J., Hopkins, T.L., 1988. Escape-
identification of zooplankton and his constructive
ment by fishes from midwater trawls: a case study using
comments on an earlier version of this manuscript. lanternfishes (Pisces: Myctophidae). Fishery Bulletin 87,
We also thank Captain Pahl and the crew of the RV 213–222.
‘Polarstern’ for outstanding work at sea. Gaskett, A.C., Bulman, C.M., He, X., Goldsworthym, S.D.,
2001. Diet composition and guild structure of mesopelagic
and bathypelagic fishes near Macquarie Island, Australia.
References NewZealand Journal of Marine Freshwater Research 35,
469–476.
Ascenio, G., Moreno, C.A., 1984. Dieta y selectividad George, E.L., Hadley, W.F., 1979. Food and habitat partiton-
alimentaria de Protomyctophum bolini Fraser-Brunner ing between Rock Bass (Ambloplites rupestris) and small-
(Pisces; Myctophidae) en el paso Drake (Antarctica). Serie mouth Bass (Micropterus dolomieui) young of the year.
cientifica. Instituto Antartico chileno. 31, 85–91. Transactions of the American Fisheries Society 108,
Barlow, K.E., Boyd, I.L., Croxall, J.P., Reid, K., Staniland, 253–261.
I.J., Brierley, A.S., 2002. Are penguins and seals in Gerasimova, O.V., 1990. Feeding and food intake of Electrona
competition for Antarctic krill at South Georgia? Marine carlsbergi (Taning, 1932) Myctophidae. Selenium Science
Biology 140, 205–213. Paper CCAMLR 7, 411–416.
Bulman, C.M., He, X., Koslow, J.A., 2002. Trophic ecology of Gon, O., Heemstra, P.C., 1990. Fishes of the Southern Ocean.
the mid-slope demersal fish community off southern J.L.B. Smith Institute of Ichthyology, Grahamstone
Tasmania, Australia. Marine Freshwater Research 53, (462pp).
59–72. Greely, T.M., Gartner Jr., J.V., Torres, J.J., 1999. Age and
Cherel, Y., Guinet, C., Tremblay, Y., 1997. Fish prey of growth of Electrona antarctica (Pisces: Myctophidae), the
Antarctic fur seals Arctocephalus gazella at Ile de Croy, dominant mesopelagic fish of the Southern Ocean. Marine
Kerguelen. Polar Biology 17, 87–90.
Biology 133, 145–158.
Clarke, K.R., Ainsworth, M., 1993. A method of linking
Groeger, J., Piatkowski, U., Heinemann, H., 2000. Beak length
multivariate community structure to environmental vari-
analysis of the Southern Ocean squid Psychroteuthis
ables. Marine Ecology-Progress Series 92, 205–219.
glacialis (Cephalopoda: Psychroteuthidae) and its use for
Clarke, K.R., Warwick, R.M., 1994. Change in Marine
size and biomass estimation. Polar Biology 23, 70–74.
Communities: An Approach to Statistical Analysis and
Guinet, C., Cherel, Y., Ridoux, V., Jouventin, P., 1996.
Interpretation. Natural Environmental Research Council,
Consumption of marine resources by seabirds and seals in
Plymouth (p. 144).
Crozet and Kerguelen waters: changes in relation to
Clarke, K.R., Warwick, R.M., 2001. Primer-E (5) Computer
consumer biomass 1962–1985. Antarctic Science 8, 23–30.
Program. Natural Environmental Reasearch Council,
Hempel, G., 1985. Antarctic marine food webs. In: Siegfried,
Plymouth.
W.R., Condy, P.R., Laws, R.M. (Eds.), Antarctic Nutrient
Clarke, T.A., 1978. Diel feeding patterns of 16 species of
Cycles. Springer, Berlin, pp. 266–276.
mesopelagic fishes from Hawaiian waters. Fishery Bulletin
Hernandez-Leon, S., Portillo-Hahnefeld, A., Almeida, C.,
76, 495–513.
Becognee, P., Moreno, I., 2001. Diel feeding behaviour of
Dalpadado, P., Gjoesaeter, J., 1988. Feeding ecology of the
krill in the Gerlache Strait, Antarctica. Marine Ecology-
lanternfish Benthosema pterotum from the Indian Ocean.
Progress Series 223, 235–242.
Marine Biology 99, 555–567.
Hoddell, R.J., Crossley, A.C., Williams, R., Hosie, G.W., 2000.
Duhamel, G., 1998. The pelagic fish community of the Polar
The distribution of Antarctic pelagic fish and larvae
frontal zone of the Kerguelen Islands. In: Di Prisco, G.,
(CCAMLR division 58.4.1). Deep-Sea Research Part II
Pisano, E., Clarke, A. (Eds.), Fishes of Antarctica: A
47, 2519–2541.
Biological Overview. Springer-Verlag Italia, Milano (Italy),
Hofmann, E.E., Klinck, J.M., Lascara, C.M., Smith, D.A.,
pp. 63–74.
1996. Water mass distribution and circulation west of the
Duhamel, G., Koubbi, P., Ravier, C., 2000. Day and night
Antarctic Peninsula and including Bransfield Strait. Ant-
mesopelagic fish assemblage off the Kerguelen Islands
(Southern Ocean). Polar Biology 23, 106–112. arctic Research Series 70, 61–80.
ARTICLE IN PRESS
C. Pusch et al. / Deep-Sea Research I 51 (2004) 1685–1708 1707
Lubimova, T.G., Shust, K.V., Troyanovskij, F.M., Semenov,
Hopkins, T.L., Gartner Jr., J.V., 1992. Resource partitioning
A.B., 1983. The ecology of mass species of myctophids from
and predation impact of a low latitude myctophid commu-
Atlantic Antarctica. Antarktika 22, 99–106.
nity. Marine Biology 114, 185–197.
McGinnis, R.F., 1982. Biogeography of laternfishes (Mycto-
Hopkins, T.L., Lancraft, T.M., 1984. The composition and
phidae) south of 301. Antarctic Research Series 35, 1–110.
standing stock of mesopelagic micronekton at 271N 861W in
Mizdalski, E., 1988. Weight and length data of zooplankton in
the eastern Gulf of Mexico. Contributions to Marine
the Weddell Sea in austral spring 1986 (ANTV/3). Berichte
Science 27, 143–158.
zur Polarforschung 55, 1–72.
Hulley, P.A., 1981. Results of the research cruises of FRV
Moku, M., Kawaguchi, K., Watanabe, H., Ohno, A., 2000.
‘‘Walther Herwig’’ to South America LVIII. Family
Feeding habits of three dominant myctophid fishes, Diaphus
Myctophidae (Osteichthyes, Myctophiformes). Archiv fur ¨
theta, Stenobrachius leucopsarus and S. nannochir, in the
Fischereiwissenschaft 31, 1–300.
subarctic and transitional waters of the western North
Hulley, P.A., 1998. Preliminary investigations on the evolution
Pacific. Marine Ecology-Progress Series 207, 129–140.
of the tribe Electronini (Myctophiformes, Myctophidae). In:
Nakamura, I., 1990. Gempylidae. In: Gon, O., Heemstra, P.C.
Di Prisco, G., Pisano, E., Clarke, A. (Eds.), Fishes of
(Eds.), Fishes of the Southern Ocean. JLB Smith Institute of
Antarctica: a biological overview. Springer-Verlag Italia,
Ichthyology, Grahamstone, South Africa, pp. 402–403.
Milano, pp. 75–85.
Naumov, A.G., 1985. System analysis of the structure of the
Hulley, P.A., Kock, K.H., Pusch, C., White, M.G., 1998.
Antarctic waters community. In: Vinogradov, M.E., Flint,
Mesopelagic fishes. Berichte zur Polarforschung 274, 17–22.
Huntley, M.E., Escritor, F., 1992. Ecology of Metridia gerlachei M.V. (Eds.), Biological Basis of the Commercial Utilization
of the Open Regions of the Ocean. Nauka Press, Moscow,
Giesbrecht in the western Bransfield Strait, Antarctica.
pp. 57–80 (in Russian).
Deep-Sea Research A 39, 1027–1055.
Pakhomov, E.A., Perissinotto, R., McQuaid, C.D., 1996. Prey
Hyslop, E.J., 1980. Stomach contents analysis-a review of
composition and daily rations of myctophid fishes in the
methods and their application. Journal of Fish Biology 17,
Southern Ocean. Marine Ecology-Progress Series 134, 1–14.
411–429.
Pearcy, W.G., Lorz, H.V., Peterson, W., 1979. Comparision of
Ichii, T., Naganobu, M., Ogishima, T., 1996. Competition
the feeding habits of migratory and non-migratory Steno-
between the krill fishery and penguins in the South Shetland
brachius leucopsarus (Myctophydae). Marine Biology 51,
Islands. Polar Biology 16, 63–70.
1–8.
Ichii, T., Katayama, K., Obitsu, N., Ishii, H., Naganobu, M.,
1998. Occurrence of Antarctic krill (Euphausia superba) Phillips, K.L., Jackson, G.D., Nichols, P.D., 2001. Predation
on myctophids by the squid Moroteuthis ingens around
concentrations in the vicinity of the South Shetland Islands:
Macquarie and Heard Islands: stomach contents and fatty
relationship to environmental parameters. Deep-Sea Re-
acid analysis. Marine Ecology-Progress Series 215, 179–189.
search I 45, 1235–1262.
Piatkowski, U., 1989. Macroplankton communities in Antarctic
Kinzer, J., 1977. Observations on feeding habits of the
mesopelagic fish Benthosema glaciale (Myctophidae) off surface waters: spatial changes related to hydrography.
Marine Ecology-Progress Series 55, 251–259.
NW Africa. In: Anderson, N.R., Zahuranec, B.J. (Eds.),
Piatkowski, U., Rodhouse, P.G., White, M.G., Bone, D.G.,
Oceanic Sound Scattering Prediction. Plenum Press, New
Symon, C., 1994. Nekton community of the Scotia Sea as
York, pp. 381–392.
sampled by the RMT 25 during austral summer. Marine
Kinzer, J., Schulz, K., 1985. Vertical distribution and feeding
Ecology-Progress Series 112, 13–28.
patterns of midwater fish in the central equatorial Atlantic.
Pielou, E.C., 1975. Ecological Diversity. Wiley, New York
I. Myctophidae. Marine Biology 85, 313–322.
(165pp).
Kock, K.-H., 1992. Antarctic Fish and Fisheries. Cambridge
Rembiszewski, J.M., Krzeptowski, M., Linkowski, T.P., 1978.
University Press, Cambridge (359pp).
Fishes (Pisces) as by-catch in fisheries of krill Euphausia
Kock, K.-H., Dollefeld, L., Hulley, P.A., Jahrig, H.-P., Petzel,
¨ ¨
superba Dana (Euphausiacea, Crustacea). Polskie archiwum
W., Pusch, C., Siegel, V., White, M.G., 1998. The
hydrobiologii 25, 677–695.
composition of the fish fauna on the northern slope of
Rodhouse, P.G., White, M.G., Jones, M.R.R., 1992. Trophic
King George Island. Berichte zur Polarforschung 274,
relations of the cephalopod Martialia hyadesi (Teuthoidea:
14–17.
Ommastrephidae) at the Antarctic Polar Front, Scotia Sea.
Kozlov, A.N., Tarverdiyeva, M.I., 1989. Feeding of different
Marine Biology 114, 415–421.
species of Myctophidae in different parts of the Southern
Rowedder, U., 1979a. Feeding ecology of the myctophid
Ocean. Journal of Ichthyology 29, 160–167.
Electrona antarctica (Guenther, 1878) (Teleostei). Meeres-
Lancraft, T.M., Torres, J.J., Hopkins, T.L., 1989. Micronekton
forschung 27, 252–263.
and Macrozooplankton in the open waters near Antarctic
Rowedder, U., 1979b. Some aspects of the biology of Electrona
Ice Edge Zones (AMERIEZ). Polar Biology 9, 225–233.
antarctica (Guenther, 1878) (Family Myctophidae). Meer-
Lancraft, T.M., Hopkins, T.L., Torres, J.J., Donnelly, J., 1991.
esforschung 27, 244–251.
Oceanic micronecton/macrozooplanktonic community
Schlitzer, R., 2003. Ocean data view. http://www.awi-bremer-
structure and feeding in ice covered Antarctic waters during
haven.de/GEO/ODV.
the winter (AMERIEZ 1988). Polar Biology 11, 157–167.
ARTICLE IN PRESS
C. Pusch et al. / Deep-Sea Research I 51 (2004) 1685–1708
1708
Shannon, C.E., Weaver, W., 1949. The Mathematical Theory of Torres, J.J., Somero, G.N., 1988. Vertical distribution and
Communication. University Illinois Press, Chicago (p. 111). metabolism in Antarctic mesopelagic fishes. Comparative
Siegel, V., 1985. The distribution pattern of krill, Euphausia Biochemistry and Physiology 90B, 521–528.
superba, west of the Antarctic Peninsula in February 1982. Tyler Jr., H.R., Pearcy, W.G., 1975. The feeding habits of three
Meeresforschung 30, 292–305. species of laternfishes (Family Myctophidae) off Oregon,
Siegel, V., de la Mare, W.K., Loeb, V., 1997. Long-term USA. Marine Biology 32, 9–11.
monitoring of krill recruitment and abundance indices in the Warwick, R.M., Clarke, K.R., 1995. New ‘biodiversity’
Elephant Island area (Antarctic Peninsula). CCAMLR measures reveal a decrease in taxonomic distinctness with
Science 4, 19–35. increasing stress. Marine Ecology-Progress Series 129,
Siegel, V., Kock, K.H., Petzel, W., Hulley, P.A., Pusch, C.,
301–305.
Dollefeld, L., White, M.G., 1998a. Investigation on the krill
¨
Williams, A., Koslov, J.A., Terauds, A., Haskard, K., 2001.
stock and the epipelagic zooplankton community in the
Feeding ecology of five fishes from the mid-slope micro-
Elephant Island region. Berichte zur Polarforschung 274,
nekton community off southern Tasmania, Australia.
50–56.
Marine Biology 139, 1177–1192.
Siegel, V., Loeb, V., Groeger, J., 1998b. Krill (Euphausia
Williams, R., 1985. Trophic relationships between pelagic
superba) density, proportional and absolute recruitment and
fish and euphausiids in Antarctic waters. In: Siegfried,
biomass in the Elephant Island region (Antarctic Peninsula)
W.R., Condy, P.R., Laws, R.M. (Eds.), Antarctic
during the period 1977–1997. Polar Biology 19, 393–398.
Nutrient Cycles and Food Webs. Springer, Berlin,
Stein, M., Heywood, R.B., 1994. Antarctic environment-
pp. 452–459.
physical oceanography: the Antarctic Peninsula and South-
Young, J.W., Blaber, S.J.M., 1986. Feeding ecology of three
west Atlantic region of the Southern Ocean. In: El-Sayed,
species of midwater fishes associated with the continental
S.Z. (Ed.), Southern Ocean Ecology—The Biomass Per-
slope of eastern Tasmania, Australia. Marine Biology 93,
spective. Cambridge University Press, Cambridge, pp.
11–24. 147–156.
Deep-Sea Research I 51 (2004) 1685–1708
www.elsevier.com/locate/dsr
Community structure and feeding ecology of mesopelagic
fishes in the slope waters of King George Island
(South Shetland Islands, Antarctica)
C. Puscha,Ã, P.A. Hulleyb, K.-H. Kockc
a
Alfred Wegener Institute for Polar and Marine Research, P.O. Box 120161, Columbusstrasse, 27568 Bremerhaven, Germany
b
Iziko Museums of Cape Town, P.O. Box 61, 8000 Cape Town, Republic of South Africa
c
Sea Fisheries Research Institute, Palmaille 9, 22767 Hamburg, Germany
Received 17 September 2003; received in revised form 18 June 2004; accepted 22 June 2004
Available online 24 August 2004
Abstract
The role of mesopelagic fishes in the Southern Ocean ecosystem and more particular their trophic effect on the
standing stock of mesozooplankton is at present poorly understood. To get a deeper insight in the Antarctic mid-water
ecosystem the mesopelagic fish community of the King George Island slope (South Shetland Islands) was sampled with
a pelagic trawl in 1996. The community structure was analysed and the feeding ecology was studied of the five most
abundant species. A total of 18 mesopelagic fish species in 10 families was identified. Of these, the Myctophidae was the
most important family by species number (9 species), individual number (98.5% of all individuals) and fish wet weight
(87.3% of the total weight). The assemblage was numerically dominated by four myctophids (Electrona antarctica,
Gymnoscopelus braueri, Gymnoscopelus nicholsi, Protomyctophum bolini) and one gempilyd (Paradiplospinus gracilis).
Multivariate statistical analysis of the mesopelagic fish data reveals two major groups of stations according to the
sampled depth: a shallow group of stations (295–450 m depth) and a deeper group of stations (440–825 m depth). The
change in relative abundance of mesopelagic fish species at 440–450 m coincides with the presence of warmer and denser
Circumpolar Deep Water at and below these depths. Deeper stations were characterized by a higher density and
increased diversity of mesopelagic fish species. The community patterns identified correlated well with the vertical depth
distribution of the most abundant species. Dietary analysis reveals that myctophids are mostly zooplanktivorous, while
the gempilyd P. gracilis is classified as a piscivorous predator. The small P. bolini feed mainly on copepods of the species
Metridia gerlachei, while the most important prey item of the larger myctophids E. antarctica, G. braueri, and G. nicholsi
were various species of euphausiids. Investigation of feeding chronology showed that G. nicholsi and P. bolini were
feeding day and night. Daily ration estimates for myctophid species ranged from 0.28% to 3.3% of dry body weight
(0.5–5.94% of wet body weight). Krill (Euphausia superba) were the most important food of E. antarctica and G.
nicholsi, accounting for 53.1% and 58.3% of the total food weight, respectively. The annual removal from the krill
ÃCorresponding author. Tel.: +49-471-4831-1652; fax:+49-471-4831-1425.
E-mail address: cpusch@awi-bremerhaven.de (C. Pusch).
0967-0637/$ - see front matter r 2004 Elsevier Ltd. All rights reserved.
doi:10.1016/j.dsr.2004.06.008
ARTICLE IN PRESS
C. Pusch et al. / Deep-Sea Research I 51 (2004) 1685–1708
1686
stock by both species was estimated to amount to 11.1–26.7% in the South Shetland Islands region. This estimate
emphasizes the important role of mesopelagic fish in the Antarctic ecosystem as a prevalent consumer of krill.
r 2004 Elsevier Ltd. All rights reserved.
Keywords: Marine fish; Myctophidae; Mesopelagic zone; Community structure; Stomach content; Predation; Daily ration
world oceans, they occupy the third level and are
1. Introduction
consumers of the second order. They are an
One-quarter of all known fish species in the important food source for the predators of higher
Southern Ocean live in the mesopelagic and trophic levels like benthopelagic fish (Bulman et
bathypelagic zones (Kock, 1992). Myctophids are al., 2002), seabirds (Guinet et al., 1996), fur seals
the dominant fish family in these zones, as far as (Cherel et al., 1997) and squid (Rodhouse et al.,
diversity, biomass and abundance are concerned. 1992; Phillips et al., 2001). Any estimation of
Thirty-three myctophid species are known from energy transport within the pelagic system must
the Southern Ocean, of which 11 have a circum- include analyses of the individual diet composition
polar distribution. Although their geographical of the mesopelagic fish and their rates of food
distribution and taxonomy have been described consumption.
(Hulley, 1981; McGinnis, 1982), comparatively Although the diet composition of the most
few studies have examined the vertical distribution abundant myctophid species is documented (Ro-
of oceanic micronekton by intensive depth strati- wedder, 1979a; Ascenio and Moreno, 1984;
fied sampling (Torres and Somero, 1988; Lancraft Williams, 1985; Kozlov and Tarverdiyeva, 1989;
et al., 1989; Piatkowski et al., 1994; Duhamel, Lancraft et al., 1991; Hoddell et al., 2000) only a
1998). All these studies indicate that the common few studies have investigated feeding chronology
Antarctic myctophids are diel vertical migrators. (Rowedder, 1979a) and estimated daily rations
King George Island is located in the southern (Gerasimova, 1990; Pakhomov et al., 1996).
Krill (Euphausia superba) plays an important
part of the Drake Passage and is strongly
influenced by the Antarctic Circumpolar Current role as the key prey item of a number of top
(ACC). The ACC is the major oceanographic predators, especially in the Atlantic Sector of the
feature of the Southern Ocean; it is an extensive Southern Ocean (Barlow et al., 2002). Because of
eastward flowing circumpolar current (Hofmann their high biomass (the total stock of the Southern
et al., 1996). The upper waters of the ACC in the Ocean was estimated by Lancraft et al. (1989) to
study area comprised Antarctic Surface Water account for 133–191 million tonnes), mesopelagic
(ASW) and the associated Circumpolar Deep fish could be one of the most important predators
Water (CDW), which flows from the Bellingshau- of oceanic zooplankton (Lancraft et al., 1989;
sen Sea into southern Drake Passage (Stein and Pakhomov et al., 1996). Numerous studies have
Heywood, 1994). The study site on the slope of shown that myctophids play a significant role in
King George Island is characterized by a shelf- the consumption of juvenile and adult krill
break front resulting in enhanced production and (Rembiszewski et al., 1978; Rowedder, 1979a;
a higher krill abundance compared to oceanic Williams, 1985). This conclusion has more recently
waters. For this reason the area north of King been challenged by a suggestion that a substantial
George Island is one of the most important krill consumption of krill occurs only during certain
fishing regions of the Southern Ocean (Ichii et al., periods and within specific regions (Pakhomov et
1996). al., 1996).
Myctophids play a significant role as consumers Finally, it should be noted that a preliminary
of zooplankton in the food web of the Southern analysis of community structure of the mesopela-
Ocean (Lancraft et al., 1989). As in the other gic fish assemblage from cruise ANT XIV/2 of RV
ARTICLE IN PRESS
C. Pusch et al. / Deep-Sea Research I 51 (2004) 1685–1708 1687
‘‘Polarstern’’ has already been presented by Hulley The sampling program consisted of 16 hauls,
et al. (1998), and the results used for a cladistic with the objective to sample three different depth
analysis of the myctophid tribe Electronini (Hul- horizons: 200–300 m, 400–500 m and near-bottom
ley, 1998). In the present study, however, commu- (2–28 m above seafloor) at soundings of 400, 600
nity structure of the mesopelagic community over and 800 m (Table 1) (Kock et al., 1998). Mesope-
the slope of King George Island is analysed in lagic fish were sampled with a pelagic trawl PT-
1088 with an estimated mouth opening of 200 m2
greater detail. The community pattern and the
vertical distribution are related to the feeding (width 20 m and a height of 10–12 m). The mesh
ecology of the most abundant mesopelagic fish size was 12 mm in the cod end. It was expected that
species. Four myctophids and one gempylid are juvenile myctophids (SLo30 mm) would not be
analysed with respect to diet, feeding chronology sampled adequately with this net configuration
and daily rations. In conclusion, an estimation is (Gartner et al., 1988). An SCANMAR depth
made of the predation impact of the mesopelagic sensor controlled the sample depth and net open-
fish community on the krill stock in the region of ing during trawling. Towing time varied between
King George Island. 30 and 60 min; trawl speeds ranged from 3.5 to
4.0 knots. Ship speed was increased during net
deployment and decreased during retrieval. This
procedure minimizes the effects of net contamina-
2. Material and methods
tion by fish resident in water layers above the
fishing depth. Station 73 was excluded from
2.1. Sampling
community analysis as the net snagged on the
bottom during trawling (Hulley et al., 1998). All
Data were collected during the cruise ANT
fishes were identified to species according to the
XIV/2 of RV ‘‘Polarstern’’ in November/Decem-
most recent keys (Gon and Heemstra, 1990). Fish
ber 1996. The study area was located over the
from the entire sample or a sub-sample of each
slope northwest of King George Island (South
species from each station were counted and
Shetland Islands), in southern Drake Passage
weighed, and standard lengths (SL) were taken
(Fig. 1).
˚S 1500 m
1000 m
61.5 500 m
51
73
70
57
52
53 68
55 69
59
58
63
64
65 54
60
200 m
orge Is.
62.0 King Ge
°W
60.0 59.5 59.0 58.5 58.0 57.5
Fig. 1. Sampling localities in 1996. Line indicates hydrographic section through the study area.
ARTICLE IN PRESS
C. Pusch et al. / Deep-Sea Research I 51 (2004) 1685–1708
1688
Table 1
Station data for 16 PT-1088 trawl samples
ST Sampling Locality Date Time Day Global Sample Bottom Ship
Timea
Longitude Latitude 1996 (Local) Radiation (wm2) Depth (m) Depth (m) Speed (km)
581430 W 611360 S
51 30 November 16:49–17:49 Day 197.8 400–450 786–875 4.0
591170 W 611450 S
52 1 December 08:00–08:30 Day 75.0 415–515 660–730 4.0
591190 W 611450 S
53 1 December 10:57–11:27 Day 153.0 273–302 635–738 4.0
591310 W 611480 S
54 1 December 14:50–15:20 Day 230.3 283–295 475–652 4.0
591340 W 611470 S
55 1 December 17:35–18:05 Day 145.8 400–450 680–995 4.0
591160 W 611430 S
57 3 December 08:19–08:50 Day 179.7 397–465 730–790 4.0
591130 W 611470 S
58 3 December 20:26–21:26 Day 12.4 283–325 380–387 3.5
591320 W 611480 S
59 3 December 23:20–00:20 Night 0.0 431–495 850–1058 3.7
591440 W 611480 S
60 4 December 02:15–03:18 Night 0.0 380–440 681–690 3.5
591320 W 611480 S
63 4 December 20:30–21:11 Day 98.5 520–580 555–608 4.0
591350 W 611470 S
64 5 December 01:06–01:41 Night 0.0 610–640 731–792 4.0
591510 W 611490 S
65 5 December 04:26–04:56 Day 80.6 750–800 1287–1468 4.0
591090 W 611440 S
68 5 December 21:10–21:41 Day 46.0 340–360 360–367 3.5
591170 W 611460 S
69 6 December 01:01–01:32 Night 0.0 560–597 584–599 3.7
591020 W 611380 S
70 6 December 04:50–05:20 Day 23.0 790–825 810–833 3.7
581440 W 611360 S
73 6 December 21:07–21:10 Day 20.0 550–575 570–580 3.5
a
Defined by sunset 22:06 and sunrise 03:23.
to the nearest millimetre with sliding callipers. (DWi ) were calculated (George and Hadley, 1979).
Hydrographic data were collected by conductivity, By the following equation, all three indices were
temperature and depth casts (CTD, 22 stations). combined to describe the prey utilization by the
The CTD was deployed at each station in advance ‘Relative Importance Index’ (RI) for each prey
category i (George and Hadley, 1979; Hyslop,
of the trawl tows.
1980):
2.2. Diet analysis ðFi þ Ni þ DWiÞ Â 100
RI i ¼ Ps ; ð1Þ
i¼1 ðFi þ Ni þ DWiÞ
Diet analysis was performed on the five most
abundant mesopelagic fish species. A maximum of
where s is the number of prey categories.
20 individuals of these species was selected from
Feeding chronology was analysed by the Sto-
each sample. In samples containing 420 speci-
mach Content Index (SCI):
mens individuals were chosen haphazardly. Fish
dry weight of stomach content
were weighed wet, measured (SL, mm below) and SCIð%Þ ¼ Â 100:
body dry weight
the whole stomach removed. The dry weight of fish
specimens was determined by oven-drying speci- ð2Þ
mens at 80 1C until constant weight was reached.
Prey organisms were identified to the lowest In addition, the stage of digestion of each prey
possible taxon and measured under a binocular item was determined by the modified method of
microscope. Dry weight of the prey organisms was Pearcy et al. (1979): Stage 1=undigested prey,
reconstructed by length–weight regressions taken Stage 2=slightly digested with some appendages
from the literature (Mizdalski, 1988; Groeger damaged, but body shape still preserved, Stage 3:
et al., 2000). body shape of prey deformed. The ratio of
Three indices, the frequency of occurrence (F i ) digestion stages was calculated for each time
of each prey item in non-empty stomachs, the interval based on these criteria.
Daily ration (mean daily food consumption, C w )
percentage of each food item by number (N i ) to
the total number, and the percentage by dry weight of the four myctophids was investigated by the
ARTICLE IN PRESS
C. Pusch et al. / Deep-Sea Research I 51 (2004) 1685–1708 1689
method of Eggers (1977): intensity were log transformed to validate the use
of normalized Euclidean distance for the calcula-
C w ¼ I Â 24=T; ð3Þ
tion of the abiotic similarity matrix. We applied
where I is the daily average SCI (%) and T the gut the Spearman Rank correlation to relate the biotic
passage time (h). matrix based on mesopelagic fish abundances with
In this study, no gut passage time data were the abiotic similarity matrix (Clarke and Ains-
collected. Therefore, data from the literature were worth, 1993).
used. Two estimates of egestion times for Antarc- Two different water masses were discernible
tic myctophids are available: the first estimate by over the slope of King George Island, separated at
Rowedder (1979a) for Electrona antarctica and the a depth of 450 m (see Section 3). Mesopelagic fish
second by Gerasimova (1990) for E. carlsbergi. samples above and below this depth were com-
These studies estimated the egestion time to be 8 pared with different sub-routines of the Primer
and 8.5 h, respectively. In our study, we used the computer program.
8.5 h suggested by Gerasimova. For the calcula- One-way analysis of similarity, ANOSIM
tion of daily ration some authors (Pakhomov et (Clarke and Warwick, 1994), was employed to test
al., 1996) have recommended the substitution of 24 the hypothesis of no differences in mesopelagic fish
 10 in Eq. (3) for species that have an active assemblage above and below 450 m. This subrou-
feeding period of 10 h. Other authors argue that tine compares the average rank similarities within
this substitution introduces a significant conserva- the predefined groups of samples with the average
tive bias (Williams et al., 2001). We calculated two similarity between groups. Values close to 1 indicate
alternative daily rations using both 10 and 24. a strong separation between groups, while a value
of 0 indicates no differences between groups.
2.3. Data analysis The similarity percentage routine (SIMPER)
(Clarke and Warwick, 1994), which was applied to
Density and biomass data for mesopelagic fish square-root transformed mesopelagic fish abun-
were calculated as individuals per filtered water dances, identified the contribution from individual
volume. The filtered volume was calculated by species to the dissimilarities between (the deep and
multiplying the trawled distance of the vessel with the shallow) sample groups.
the estimated mouth opening (200 m2) of the PT- Various univariate indices were calculated in
1088. order to characterize the species assemblages of the
Community structure was investigated with the deep and shallow group of samples: species
number, Shannon’s diversity index (H0 ) (Shannon
Primer-E5 Software package (Clarke and War-
wick, 2001). To reduce the weighting of dominant and Weaver, 1949) and Pielou’s Evenness Index (J)
species, the densities were square-root transformed (Pielou, 1975). As these indices are known to be
prior to the computation of the triangular influenced by sample size, we also calculated the
similarity matrices based on Bray-Curtis simila- taxonomic diversity D and taxonomic distinctness
DÃ ; which consider the taxonomic relatedness of
rities (Field et al., 1982). The results of the latter
were classified by hierarchical agglomerative clus- species (Warwick and Clarke, 1995). Taxonomic
diversity D is empirically related to H 0 but
ter analysis using the group average linking
method, and ordinated by a non-metric, multi- contains, in addition, information on the taxo-
dimensional scaling technique (MDS). nomic separation of the species in a sample, i.e.
The BIOENV sub-routine was used to relate the besides the distribution of individuals among
community patterns of the mesopelagic fish species it also takes into account the distribution
assemblage to six environmental variables: light of species in the taxonomic system by weighting
intensity (W/m2) (indicating time of day), mini- the co-occurrences of species according to the
mum and maximum values of sample depth, degree of separation in the hierarchical classifica-
bottom depth, temperature and salinity at the tion (1=different species, 2=different genera,
3=different families, 4=different orders). DÃ is
sampled depth horizon. Values of average light
ARTICLE IN PRESS
C. Pusch et al. / Deep-Sea Research I 51 (2004) 1685–1708
1690
derived from D but measures solely the taxonomic Sea and Scotia Sea mingle. Analysis of vertical
distinctness of species in a sample, without the temperature and salinity sections through the
contribution from species diversity. study site revealed the existence of two different
Oceanographic data were obtained from the water masses in the slope region of King
program Ocean Data View (Schlitzer, 2003). The George Island (Fig. 2A–B). The water mass in
results are presented as a section through the study the upper 200 m was composed of ASW, char-
area. acterized by a temperature minimum in 50–150 m
depth (To0:5 1CÞ and a low salinity (o 34.4). The
origin of this water body is due to the cooling of
the surface water during winter. The properties of
3. Results
this so-called winter water are stable year round
(Hofmann et al., 1996). During our study the
3.1. Oceanographic conditions
surface water temperature was slightly
increased by enhanced solar radiation. The
The study area is situated in the Weddell-Scotia-
dominating water mass below 400 m depth,
Confluence, where water masses of the Weddell
Fig. 2. (A–B) RV ‘‘Polarstern’’ cruise ANT XIV/2, east–west hydrographic section through the study area as indicated in Fig. 1, (A)
temperature (1C) and (B) salinity.
ARTICLE IN PRESS
C. Pusch et al. / Deep-Sea Research I 51 (2004) 1685–1708 1691
Circumpolar Deep Water, was characterized by other families represented by just one species.
Three myctophids, Gymnoscopelus fraseri, G.
higher temperature (T40.5 1C) and salinity
hintonoides and Protomyctophum choriodon, were
(S434:5). A transition zone between these water
masses occupies the depth range 200–400 m, in newly recorded for the South Shetland Island
which the gradients of temperature and salinity region.
were strongest. Myctophids dominated the mesopelagic fish
community by number and wet weight accounting
for 98.5% of all sampled individuals and 87.3% of
3.2. Species composition
the total weight. E. antarctica was the most
abundant species by number and weight. This
The total catch of mesopelagic fish taken on 15
myctophid accounted for 60.7% of all sampled
pelagic trawl stations included 16 343 specimens
individuals, followed by P. bolini (19.6%), G.
with a wet weight of 122.5 kg (Table 2). Eighteen
braueri (13.8%) and G. nicholsi (4.0%). All other
mesopelagic species belonging to 10 families and
mesopelagic fish species were of numerically minor
13 genera were identified. The family Myctophidae
importance (o2% in total).
was by far the most speciose (9 species), with all
Table 2
Number of individual fish, density as number of individuals per 105 m3 filtered volume, wet weight in kg and biomass as g wet weight
per filtered volume in 15 PT-1088 samples
Species Abundance Biomass
Ind/105 m3 g ww/105 m3
No. Inds. (%) kg (%)
Astroneshtidae
Borostomias antarcticus 1 (0.01) 0.00 0.01 (0.01) 0.05
Bathylagidae
Bathylagus antarcticus 4 (0.02) 0.03 0.05 (0.04) 0.43
Chiasmodontidae
Chiasmodon niger 1 (0.01) 0.00 0.02 (0.02) 0.09
Gempylidae
Paradiplospinus gracilis 182 (1.11) 1.14 14.96 (12.21) 93.29
Gonostomatidae
Cyclothone sp. 2 (0.01) 0.02 0.01 (0.01) 0.09
Melamphaidae
Poromitra crassiceps 2 (0.01) 0.01 0.02 (0.02) 0.09
Microstomatidae
Nansenia antarctica 3 (0.02) 0.02 0.09 (0.07) 0.66
Myctophidae
Electrona antarctica 9931 (60.71) 59.48 67.93 (55.46) 407.98
Gymnoscopelus braueri 2253 (13.77) 20.23 13.71 (11.19) 122.72
Gymnoscopelus fraseri 2 (0.01) 0.01 0.02 (0.02) 0.14
Gymnoscopelus hintonoides 15 (0.09) 0.11 0.13 (0.11) 0.99
Gymnoscopelus nicholsi 647 (3.96) 4.2 21.15 (17.27) 136.81
Gymnoscopelus opisthopterus 27 (0.17) 0.13 0.96 (0.78) 4.52
Krefftichthys anderssoni 23 (0.14) 0.18 0.07 (0.06) 0.55
Protomyctophum bolini 3212 (19.63) 20.1 2.97 (2.42) 18.65
Protomyctophum choriodon 1 (0.01) 0.00 0.01 (0.01) 0.04
Notosudidae
Scopelosaurus hamiltoni 3 (0.02) 0.02 0.27 (0.22) 1.48
Paralepididae
Notolepis coatsi 34 (0.21) 0.22 0.11 (0.09) 0.71
Total 16 343 122.5
ARTICLE IN PRESS
C. Pusch et al. / Deep-Sea Research I 51 (2004) 1685–1708
1692
In wet weight, E. antarctica accounted for made. Nevertheless, some assumptions about the
55.6% of total mesopelagic fish biomass, far vertical distribution patterns of the more abundant
exceeding G. nicholsi (17.3%), G. braueri (11.2) species are possible. All four myctophids were
and the gempilyd, Paradiplosinus gracilis (12.2%). more abundant, in the 273–825 m depth range, in
daytime samples than in nighttime samples (Fig.
3A–D). This observation supports the results of
3.3. Patterns of vertical distribution
earlier studies that the bulk of Antarctic mycto-
phids migrate to epipelagic layers at night (Torres
Because of the sampling strategy (no samples
and Somero, 1988; Lancraft et al., 1989).
above 273 m depth and only four night tows) no
P. gracilis was concentrated in depths below
detailed conclusions about the vertical migration
400 m during daytime (Fig. 3E). The distribution
behaviour of the mesopelagic community can be
(B) G. braueri
0
0 (A) E. antarctica
200
200
400
400
600
600
800
800
300 200 100 0 100 200 300
300 200 100 0 100 200 300
(D)
0 0
(C)
G. nicholsi P. bolini
200 200
Depth (m)
400 400
600 600
800 800
20 10 10 20 80 40 0 40 80
0
(E)
0
P. gracilis
200
400
600
800
4
4 2 0 2
Density
Fig. 3. (A–E) Vertical day/night distribution: (A) E. antarctica, (B) G. braueri, (C) G. nicholsi, (D) P. bolini, (E) P. gracilis, Density as
individuals per 105 m3; open bars=day tows; filled bars=night tows.
ARTICLE IN PRESS
C. Pusch et al. / Deep-Sea Research I 51 (2004) 1685–1708 1693
data show an increased density in this species
between 400 and 500 m, suggesting vertical migra-
tion behaviour at least in mesopelagic depths. The
four myctophids showed different daytime depth
distributions: The bulk of E. antarctica individuals
were found below 400 m depth (Fig. 3A) and the
minimum depth of occurrence of G. braueri was
750 m (Fig. 3B). In contrast, P. bolini occurred
over a depth range of 273–825 m, with the bulk of
the population in the shallow 273–450 m interval
(Fig. 3D). G. nicholsi occurred over the whole Fig. 4. Dendrogram representing the classification of 15 PT-
1088 stations according to mesopelagic fish assemblage; Station
sampled depth range (273–825 m) with a centre of
number and maximum sampled depth are indicated. Hatching
distribution between 400 and 597 m (Fig. 3C).
indicates shallow stations (depth 302–450 m).
3.4. Community analysis
Deep
Shallow
Cluster analysis based on densities of mesope-
lagic fish showed a clear separation of stations at 68
an arbitrary level of 40% (Fig. 4). The first cluster
65
comprises stations sampled at and above 450 m 70
depth. The second cluster is exclusively composed 58 54
60 59
69
of stations taken in depth4450 m, with the 52
53 55
57
exception of station 60, where sampling depth 63
was 380–440 m. 64
The separation of stations at 450 m depth is
confirmed by the ordination with MDS of the
51
same assemblage data (Fig. 5). Again, station 60
was more closely associated with the deeper group
of stations. Fig. 5. MDS plot for 15 PT-1088 stations according to the
mesopelagic fish assemblage; shallow and deep refer to the
An exploratory analysis, BIOENV, was applied
sampled depth; shallow=depth 302–450 m; deep=depth
to examine which abiotic variables could best
465–825 m; stress (=goodness of fit)=0.07.
explain the observed patterns in the mesopelagic
fish assemblage. The maximum matching coeffi-
cient was achieved by the abiotic variable max- An analysis of similarity (ANOSIM) was
imum sample depth (51.7%) and supported performed to test for statistical differences in
therefore the result from the MDS and cluster species composition between shallow stations
analyses. Minimum sample depth (46.4%) and (295–450 m depth) and deep stations
(465–825 m). A value of R ¼ 0:561 supported the
minimum and maximum salinity (34.9% and
28.6%, respectively) were further useful abiotic results of the classification and ordination of the
parameters to explain the observed pattern in the data and indicated significant differences in species
mesopelagic fish community. Other abiotic vari- composition between shallow and deep stations
ables showed a low correlation with the biotic (P ¼ 0:003).
similarity matrix and yielded matching coefficients The similarity percentage procedure (SIMPER)
below 25%. Sample depth is thus the best was applied to identify those species that con-
environmental variable to explain the grouping tribute most to the observed differences between
of the samples in a manner consistent with the shallow and deep samples (Table 3). Only three
biotic pattern. species were more abundant in the shallow group
ARTICLE IN PRESS
C. Pusch et al. / Deep-Sea Research I 51 (2004) 1685–1708
1694
Table 3
SIMPER analysis; discriminating species between ‘‘deeper’’ and ‘‘shallower’’ station groups as revealed by cluster analysis
Avg. density (Ind/105 m3)
Species Cumulative
Shallow Deep Contribution (%) Contribution (%)
Electrona antarctica 19.01 94.89 43.33 43.33
Protomyctophum bolini 33.6 8.3 18.8 62.13
Gymnoscopelus braueri 0.01 37.93 16.35 78.48
Gymnoscopelus nicholsi 4.13 4.26 5.89 84.36
Paradiplospinus gracilis 0.3 1.88 5.22 89.59
Notolepis coatsi 0.05 0.37 2.29 91.88
Krefftichthys anderssoni 0 0.34 2.28 94.16
Gymnoscopelus hintonoides 0 0.2 1.67 95.82
Gymnoscopelus opisthopterus 0 0.24 1.19 97.02
Bathylagus antarcticus 0 0.06 0.5 97.52
Nansenia antarctica 0.01 0.03 0.47 97.99
Scopelosaurus hamiltoni 0 0.03 0.46 98.45
Gymnoscopelus fraseri 0 0.03 0.41 98.85
Protomyctophum choriodon 0 0.02 0.28 99.13
Poromitra crassiceps 0.01 0 0.25 99.38
Cyclothone sp. 0 0.04 0.25 99.62
Chiasmodon niger 0 0.01 0.2 99.82
Borostomias antarcticus 0.01 0 0.18 100
Species are ordered in decreasing contribution (%) to the total dissimilarity.
Table 4
of samples, while 15 species were more abundant
Univariate indices of the deeper and shallower groups of
in the deeper group. The myctophid P. bolini was
stations over the slope of King–George Island
the best indicator species for samples taken above
450 m depth and accounted for 18.8% of the Shallow Deep
observed differences. Deeper stations were char-
No. of species 9 16
acterized by relatively higher abundances of E.
Density (Ind./105 m3) 57.13 148.61
antarctica and G. braueri and P. gracilis , which Pielou’s evenness (J) 0.35 0.41
together accounted for 65% of the observed Shannon’s diversity (H0 ) 0.51 0.79
differences between the two groups of samples. Taxonomic diversity (D) 10.03 14.64
Taxonomic distinctness (DÃ ) 35.87 37.02
The calculation of several univariate indices for
the species densities above and below 450 m depth
reflected the observed differences between the
not significant. The taxonomic distinctness DÃ
mesopelagic fish assemblages (Table 4). The
number of species (9) in depth between 273 and showed a high similarity between shallow and
450 m was lower, compared to 16 species in depths deep samples.
below 450 m (two individuals of Cyclothone sp.
3.5. Feeding ecology
were not identified to species level). Also the
average density of mesopelagic fish was 2.5 times
3.5.1. Food composition
higher in the deeper group samples. Shannon’s
diversity (H0 ), Pielou’s evenness index (J) and the The diet of E. antarctica was dominated by
taxonomic diversity D were higher in samples euphausiids in our samples. By number, euphau-
taken below 450 m depth. However, the observed siids (47.8%) are followed by ostracods (15.6%)
differences in diversity indices between the deeper and then by copepods (26.7%) (Fig. 6A; Table 5).
and shallower group of samples were statistically By dry weight (DW), euphausiids account for
ARTICLE IN PRESS
C. Pusch et al. / Deep-Sea Research I 51 (2004) 1685–1708 1695
G. braueri (n = 21)
E. antarctica (n = 61)
DW (%) A (%)
DW (%) N (%)
100 80 60 40 20 0 20 40 60 80 100
100 80 60 40 20 0 20 40 60 80 100
20 20
Eu (RI = 61.1) Eu (RI = 52.9)
40 40
60 60
Cop (14.9)
80 80 Os (19.0)
Os (12.0)
100 100
Co (13.9)
Am (6.0)
Ch (4.3)
Am (9.3)
Ch (4.9)
F (%) (B)
(A) F (%)
G. nicholsi (n = 135) P. bolini (n = 101)
DW (%) N (%) DW (%) N (%)
100 80 60 40 20 0 20 40 60 80 100 100 80 60 40 20 0 20 40 60 80 100
20 20
Eu (RI = 44.0)
40 40 Co (RI = 82.0)
60 60
80 80
Co (37.9)
100 100 Os (10.0)
Eu (6.2)
Os (7.3)
F (%)
(D)
(C)
Ch (5.7)
Pi (4.7)
F (%)
P. gracilis (n = 28)
DW (%) N (%)
100 80 60 40 20 0 20 40 60 80 100
Am: Amphipods
Ceph: Cephalopds
20
Ch: Chaetognaths
Pi (RI = 76.8) 40
Co : Copepods
60
Eu : Euphausiids
80
Os : Ostracods
Eu (15.1)
100
Pi : Pisces
Ceph (8.4)
(E) F (%)
Fig. 6. (A–E) Percentage composition of the main prey items of (A) E. antarctica (B) G. braueri (C) G. nicholsi (D) P. bolini (E) P.
gracilis, by percent dry weight (DW), percentage number (N) and frequency of occurrence (F), The relative importance (RI) Index is
presented by the size of the respective rectangles. n=number of stomachs containing food.
1696
Table 5
Diet composition of E. antarctica, G. braueri, G. nicholsi, P. bolini and P. gracilis, showing number and reconstructed dry weight of each food item; the respective
percentage of the total number and the total dry weight is given in parenthesis; ‘‘-’’: absent.
Electrona antarctica Gymnoscopelus braueri Gymnoscopelus nicholsi Protomyctophum bolini Paradiplospinus gracilis
SL range (mm) 57–113 69–121 123–172 33–53 284–469
No. of fish examined 145 67 186 122 100
No. of stomach empty (%) 84 (57.9) 46 (68.7) 51 (27.4) 21 (17,2) 72 (72)
C. Pusch et al. / Deep-Sea Research I 51 (2004) 1685–1708
No. (%) Weight (%) No (%) Weight (%) No (%) Weight (%) No. (%) Weight (%) No. (%) Weight %
Amphipoda 2 (2.2) 210.86 (13.9) 1 (3.2) 84.08 (23.5) 1 (0.1) 50.75 (0.8) — — — — — — — —
Cephalopoda — — — — — — — — — — — — — — 3 (8.3) 9900 (7.4)
Psychroteuthis glacialis 3 (8.3) 9900 (7.4)
ARTICLE IN PRESS
Chaetognatha 5 (5.6) 3.9 (0.3) 2 (6.5) 1.56 (0.4) 21 (2.6) 241.78 (3.8) 4 (0.3) 3.12 (0.6) — — — —
Copepoda 24 (26.7) 5.42 (0.4) 7 (22.6) 1.97 (0.6) 517 (63.8) 235.67 (3.7) 1515 (96.6) 354.93 (84.2) — — — —
Calanus propinquus 1 (1.1) 0.8 (0.1) — — — — 23 (2.8) 21.84 (0.3) 8 (0.5) 6.21 (1.5) — — — —
Gaidius spp. 1 (1.1) 0.37 (0.0) 3 (9.7) 1.11 (0.3) 10 (1.2) 3.7 (0.1) 11 (0.7) 4.07 (1.0) — — — —
Metridia gerlachei 11 (12.2) 1.62 (0.1) — — — — 341 (42.1) 68.58 (1.1) 1438 (91.7) 319.83 (75.9) — — — —
Pareuchaeta spp. 1 (1.1) 1.35 (0.1) — — — — 67 (8.3) 95.82 (1.5) 13 (0.8) 15.87 (3.8) — — — —
Rhincalanus gigas — — — — 1 (3.2) 0.43 (0.1) 46 (5.7) 29.69 (0.5) 5 (0.3) 2.49 (0.6) — — — —
Unidentified calanoids 10 (11.1) 1.28 (0.1) 3 (9.7) 0.43 (0.1) 30 (3.7) 16.04 (0.3) 40 (2.6) 6.46 (1.5) — — — —
Euphausiacea 43 (47.8) 1292.07 (85.0) 13 (41.9) 268.66 (75.0) 226 (27.9) 4950.82 (77.9) 8 (0.5) 54.79 (13.0) 8 (22.2) 280.69 (0.2)
Euphausia cristallorophias — — — — — — — — 1 (0.1) 7.55 (0.1) — — — — — — — —
E. frigida 1 (1.1) 14.63 (1.0) — — — — 2 (0.2) 78.12 (1.2) — — — — — — — —
E. superba 14 (15.6) 807.23 (53.1) — — — — 67 (8.3) 3707.89 (58.3) — — — — 6 (16.7) 223.19 (0.2)
E. triacantha 1 (1.1) 19.96 (1.3) 1 (3.2) 36.82 (10.3) 4 (0.5) 163.74 (2.6) — — — — 1 (2.7) 45.4 (0.0)
Euphausia spp. 22 (24.4) 425.04 (28.0) 12 (38.7) 231.84 (64.8) 20 (2.5) 386.4 (6.1) — — — — — — — —
Thysanoessa macrura 5 (5.6) 25.21 (1.7) — — — — 132 (16.3) 607.12 (9.6) 8 (0.5) 54.79 (13.0) 1 (2.7) 12.1 (0.0)
Gastropoda 1 (1.1) 1.56 (0.1) — — — — — — — — 1 (0.1) 1.56 (0.4) — — — —
Ostracoda 14 (15.6) 4.78 (0.3) 8 (25.8) 1.74 (0.5) 39 (4.8) 8.76 (0.1) 41 (2.6) 7.13 (1.7) — — — —
Pisces — — — — — — — — 6 (0.7) 866.97 (13.6) — — — — 25 (69.4) 123868.28 (92.4)
Electrona antarctica — — — — — — — — — — — — — — — — 2 (5.6) 6435.36 (4.8)
Gymnoscopelus nicholsi — — — — — — — — — — — — — — — — 1 (2.7) 10162.92 (7.6)
Gymnoscopelus braueri — — — — — — — — — — — — — — — — 1 (2.7) 2270 (1.7)
Neopageotopsis ionah — — — — — — — — 2 (0.3) 408.25 (6.4) — — — — — — — —
Protomyctophum bolini — — — — — — — — 4 (0.5) 458.72 (7.2) — — — — — — — —
Polychaeta 1 (1.1) 1.84 (0.1) — — — — — — — — — — — — —
Total 90 1520.43 31 358.01 810 1742.7 1569 7.13 36 134048.97
ARTICLE IN PRESS
C. Pusch et al. / Deep-Sea Research I 51 (2004) 1685–1708 1697
85.0% of the total stomach content, far exceeding the identifiable food items respectively, while other
amphipods (13.9%) and the other prey categories food items accounted for less than 5% each. In dry
(o 2% in total). Euphausiids occurred in 65.6% of weight (DW) euphausiids made up 77.9% of the
all filled stomachs, while ostracods and copepods total prey, followed by fish (13.6%), chaetognaths
occurred less frequently (21.1% and 23.0%, (3.8%) and copepods (3.7%). The frequencies of
respectively). The RI values were 61.1% for occurrence for copepods, euphausiids, ostracods
euphausiids, 14.9% for copepods and 12.0% for and chaetognaths were 77.7%, 62.9%, 22.9% and
ostracods, stressing the importance of euphausiids 15.5%, respectively, while the remaining prey
in the diet of E. antarctica. categories accounted for less than 4.4% in total.
The size distribution of the 145 individuals The RI values were 44.0% for euphausiids and
examined for food composition was unimodal and 37.9% for copepods, while the remaining prey
ranged from 57 to 113 mm (Fig. 7A). To investi- categories accounted foro10%.
gate an ontogenetical shift in the prey size of E. The size distribution of the 186 individuals of G.
antarctica we compared the ratio of euphausiids nicholsi examined for stomach content was uni-
(largest prey item) to the total dry weight of modal and ranged from 123 to 172 mm (Fig. 7C).
stomach content in individuals smaller and larger The ratio of euphausiids (largest prey item) to the
than 80 mm SL. The ratio of euphausiids to the total dry weight of the stomach content was
54.7764.0% (avg. 7SD) for 123–144 mm SL
total dry weight of stomach contents was
60.7749.0% (avg.7SD) for 57–79 mm sized individuals and 58.3745.1% for 145–172 mm SL
individuals, and 68.6746.2% for 80–113 mm SL fish. The difference was not statistically significant
(Mann–Whitney U test, p40:05) and we therefore
fish. The difference was not significant (Man-
n–Whitney U-test, p40:05). Therefore, no signifi- suggest that no substantial ontogenetic shift in
cant shift in prey size was detectable over the size prey size occurs over the size range examined.
Prey species composition showed that G. nicholsi
range examined.
In this study, feeding intensity of E. antarctica fed on all five euphausiids species occurring in the
was low, with 57.9% of stomachs empty (Table 5). Southern Ocean (Table 5). The most important
euphausiid by food weight was E. superba (58.3%
More then half of all euphausiids were not
of total food weight), while Metridia gerlachei
identifiable to species. Nevertheless, krill (E.
superba) was the most important euphausiid in dominated numerically (42.1%). G. nicholsi was
the diet of this myctophid, accounting for more the only myctophid in our study that preyed on
other mesopelagic fish (Neopageotopsis ionah and
than one-half (53.1%) of the overall food weight.
The prey composition of G. braueri was Protomyctophum bolini).
The food spectrum of P. bolini, a relatively
similarly dominated by euphausiids (Fig. 6B,
Table 5). Euphausiids were most important by small-sized myctophid (SL 33–53 mm, Fig. 7D),
number (N ¼ 41:9%), DW (75.0% of total weight) was dominated by copepods (Fig. 6D, Table 5).
The N, DW and F values for copepods were
and frequency of occurrence (61.9%). The RI
reflected the high dietary value of euphausiids 96.6%, 84.2% and 91.1%, respectively. The RI
(52.9%), while ostracods, copepods and amphi- value of copepods was 82.0% followed by
pods were of minor importance with RI values of ostracods (10.0%) and euphausiids (6.2%).
19.0%, 13.9% and 9.3%, respectively. Despite the The prey species composition showed that
copepods of the genus M. gerlachei were by far
high significance of euphausiids in the diet, the
most abundant euphausiid, E. superba, was lack- the most important prey item, accounting for
ing (Table 5). The feeding intensity of G. braueri 91.7% of all identified food items (Table 5).
The diet of P. gracilis, the only non-myctophid
was noticeably low, with 68.7% of stomachs
containing no identifiable food. in our analysis, consisted almost exclusively of fish
The diet of G. nicholsi, was numerically domi- (Fig. 6E, Table 5). The three trophic parameters N,
DW and F for this prey category were 69.4%,
nated by copepods and euphausiids (Fig. 6C,
Table 5), accounting for 63.8% and 27.9% (N) of 92.4% and 82.1%, resulting in an RI of 76.8%.
ARTICLE IN PRESS
C. Pusch et al. / Deep-Sea Research I 51 (2004) 1685–1708
1698
(B)
(A)
100 100
E. antarctica G. braueri
n = 582 n = 187
80 80
60 60
40 40
20 20
0 0
0 20 40 60 80 100 120140 160180 200 0 20 40 60 80 100120140 160180 200
(D)
100
100 (C)
G.nicholsi P. bolini
n = 299 n = 189
80
80
Number of Fish
60
60
40
40
20
20
0
0
0 20 40 60 80 100120140 160180 200
0 20 40 60 80 100120 140160180 200
Length (SL, mm)
100 (E)
P. gracilis
n = 105
80
60
40
20
0
0 20 40 60 80 100120 140160180 200
Length (SL, cm)
Fig. 7. (A–E) Length-frequency plots of (A) E. antarctica (B) G. braueri (C) G. nicholsi (D) P. bolini (E) P. gracilis showing size classes
of fish measured (open bars) and sampled for diet analysis (filled bars); n total number of fish measured.
ARTICLE IN PRESS
C. Pusch et al. / Deep-Sea Research I 51 (2004) 1685–1708 1699
The average and range of SCI for G. nicholsi
Most fishes found in the stomachs were not
identifiable because of the advanced stage of were calculated to be 0.29% and 0.12% to 0.55%
digestion. Nevertheless, the common myctophids (Fig. 9A). There was no statistical difference in
E. antarctica, G. nicholsi and G. braueri could be feeding intensity between samples taken at day and
U-test, p ¼ 0:178).
identified. Additional euphausiids and cephalo- night (Mann–Whitney
pods (Psychroteuthis glacialis) were found in the However, feeding intensity was increased during
stomachs of P. gracilis, accounting for RI values evening times at 2100–2200 h (average SCI 0.38%)
of 15.1% and 8.4%, respectively. and reached a maximum at 2300–2400 h (SCI
0.55%). The observation of intensified feeding at
3.5.2. Feeding chronology these time periods is supported by a higher
The influence of sample depth on feeding proportion of undigested food items. We observed
intensity was investigated by the calculation of a decrease in feeding intensity after midnight and
the Stomach Content Index (SCI) for E. antarc- lasting until the early morning hours. The sto-
tica, G. nicholsi and P. bolini. We compared the machs of G. nicholsi contained fresh food items
SCIs of individuals taken in shallow (sample depth during day and night (stage 1) (Fig. 9C). For this
295–440) and deep tows (465–825 m) and found no reason we suggest that feeding activity was taking
differences for E. antarctica, G. nicholsi and P. place continuously.
bolini (Mann–Whitney U-Test; p40:05). There- In the small P. bolini average SCI (1.1%) was
higher than in G. nicholsi (Fig. 9B). We excluded
fore, it was concluded that sample depth had no
major influence on the SCI of these species. G. one station (58) with extraordinarily high
braueri was taken in too low numbers in the upper stomach filling (SCI 10.6%), and which we
depth horizon to be able to draw conclusions interpret as an outlier. No difference in median
about the influence of sample depth on feeding SCI was found between daytime and nighttime
samples (Mann–Whitney U test p ¼ 0:127). Never-
behaviour.
The SCI of E. antarctica fluctuated between theless, feeding activity was increased during hours
0.0% and 0.9% (avg. 0.47) over a 24 h time period of darkness when SCI values were above the
(Fig. 8A). Highest feeding activity was observed at average. A second peak in feeding intensity was
night (0200–0300 h). Minimum values in feeding observed in the morning hours from 0800 to
intensity were found for the intervals 0800–0900 h 0900 h (SCI 1.7%). During the day, SCI reached a
and 1100–1200 h. Nevertheless, no difference in minimum value of 0.21 at 1400–1500 h. The
feeding intensity was found between day and night highest proportions of undigested food items
samples (Mann–Whitney U-test, p ¼ 0:081), be- were recorded in the evening and before midnight
cause of the high variability of the SCI values (Fig. 9D).
during day. The proportion of undigested food
3.5.3. Daily ration and the impact of predation on
items (Stage 1) was reduced during daytime except
the krill stock
at one station during the time interval 1100–1200 h
where results are based on a low number of fish (5) The daily ration estimate was highest for the
small myctophid P. bolini, accounting for 2.48%
and food items (2) (Fig. 8C).
The average SCI of G. braueri was 0.22%, and of fish wet weight assuming a 10 h feeding period
values ranged from 0.04% to 0.4% (Fig. 8B). and 5.94% assuming that the species feeds 24 h a
It is difficult to draw conclusions about the day (Table 6). The lowest daily ration was
calculated for G. braueri, accounting for 0.5%
feeding chronology of this species because of the
low overall feeding intensity. We found no wet body weight under the conservative assump-
statistical differences in SCI between day and tion and 1.19% with a 24 h feeding period. The
night samples (Mann–Whitney U-test, p ¼ 0:867). estimated daily rations of G. nicholsi are calculated
to be only slightly higher than in G. braueri at
The proportion of fresh prey items was highest
0.65% and 1.57%. E. antarctica took an inter-
in the early evening (time interval 2000–2100 h)
(Fig. 8D). mediate position with daily rations of 1.06% and
ARTICLE IN PRESS
C. Pusch et al. / Deep-Sea Research I 51 (2004) 1685–1708
1700
Fig. 8. (A–D) Diel changes in stomach content of E. antarctica and G. braueri: (A–B) Average Stomach Content Index (SCI); vertical
bars=standard error; n=number of individuals examined. (C–D) Proportion of digestion stage (stage 1=fresh food, stage 2=slightly
digested, stage 3=only indigestible remains). Sunset (SS) and sunrise (SR) are indicted on the x-axis. Filled bars represent hours of
darkness.
2.54% of wet body weight by the 10 and 24 h region in the years 1996/1997 was estimated to be
92 g/1000 m3 (Siegel et al., 1998b). We estimated
feeding period assumption.
Two myctophids in our study, E. antarctica and the biomass of E. antarctica and G. nicholsi to be
4.08 and 1.36 g/1000 m3, respectively, in slope
G. nicholsi, fed on krill (E. superba). Based on the
food composition data and the daily rations, we waters off King George Island. The daily intake
of krill by E. antarctica was assessed to be 1.06%
assessed the predation impact of these myctophids
on the krill stock in the study area. of its body wet weight during a 10 h feeding period
E. superba was the most important prey item in and 2.54% in 24 h.
the diet of E. antarctica and G. nicholsi and The predation impact of G. nicholsi on krill was
accounted for 53.1% and 58.3% of the total food estimated to be 0.65% of body wet weight
weight in these species, respectively. The standing under the assumption of a 10 h feeding period
stock of krill biomass in the Elephant Island and 1.57% assuming a 24 h feeding period.
ARTICLE IN PRESS
C. Pusch et al. / Deep-Sea Research I 51 (2004) 1685–1708 1701
Fig. 9. (A–D) Feeding periodicity of G. nicholsi and P. bolini, same abbreviations as in Fig. 8.
Neglecting seasonal variability in consumption, micronekton community in the Ice-free Zone of
our data suggest an annual removal of 9.0–21.8% the Scotia Sea (Lancraft et al., 1989; Piatkowski
of the krill stock by E. antarctica and 2.1–5.0% by et al., 1994).
G. nicholsi. It was expected that the mesopelagic fish
assemblage at King George Island, which is
localized in the Seasonal Pack-ice Zone, would
be characterized by a lower species number and
4. Discussion
diversity compared to the Ice-free Zone (Kock,
4.1. Species composition 1992). The relatively high number of species over
the slope of King George Island could be due to
Eighteen mesopelagic fish species were identified increased productivity over the slope of King-
in samples in the slope waters of King George George Island, itself due to a shelf-break front as
Island, situated in the seasonal pack ice zone. This described by Ichii et al. (1998). Unpublished
number agrees with previous studies of the LIDAR measurements performed on cruise ANT
ARTICLE IN PRESS
C. Pusch et al. / Deep-Sea Research I 51 (2004) 1685–1708
1702
Table 6
Daily ration and predation impact on the krill stock of four myctophid species
E. antarctica G. braueri G. nicholsi P. bolini
Avg. SCI (% body dry weight) 0.47 0.22 0.29 1.10
Feeding period (h) 10/24 10/24 10/24 10/24
Daily ration (% body dry weight) 0.59/1.41 0.28/0.66 0.36/0.87 1.38/3.3
Daily ration (% body wet weight)a 1.06/2.54 0.50/1.19 0.65/1.57 2.48/5.94
Proportion of E. superba in food weight (%) 53.1 0.00 58.30 0.00
Daily ration of E. superba (% body wet weight) 0.56/1.35 0.00 0.38/0.92 0.00
Predator density (g/1000 m3) 4.08 1.20 1.36 0.19
Daily consumption of krill (mg/1000 m3) 22.85/55.08 0.00 5.17/12.51 0.00
Annual consumption of krill (g/1000 m3) 8.32/20.05 0.00 1.88/4.55 0.00
Percentage of krill stock removed (%)b 9.04/21.79 0 2.05/4.95 0
Daily rations were calculated for both 10 and 24 h periods of feeding.
a
Conversion factor daily ration dw to ww 1.8.
b
Taking in account a krill biomass of 92 g/1000 m3 (Siegel et al., 1998b).
XIV/2 indicated an increased chlorophyll-a con- Front (Rowedder, 1979b; Hulley, 1981; McGinnis,
centration in the vicinity of the Scotia-Weddell 1982; Lancraft et al., 1989; Hoddell et al., 2000).
The size spectrum of E. antarctica suggested the
Confluence, which passed through the study site.
A higher diversity of mesopelagic fish, resulting presence of the age classes two and three during
from an ‘island effect’, was described by Piat- our study, according to the results of Greely et al.
kowski et al. (1994) for an assemblage above the (1999).
The centre of distribution of P. choriodon, G.
slope of South Georgia when compared to an
fraseri and G. hintonoides are known to be in the
oceanic station in the northern Scotia Sea. The
snake mackerel P. gracilis is a characteristic area of the South Polar Front and north of it
species of slope habitats with benthopelagic (Hulley, 1981). The first records of these species in
life style (Nakamura, 1990) which was also the South Shetland Islands region give support to
recorded over the slope of South Georgia the idea that frontal zones, such as the South Polar
(Piatkowski et al., 1994). Front Zone, are not impervious barriers but may
However, our study supports the results of be penetrated by mesopelagic fish that are trapped
previous studies that characterize the Antarctic in eddies (Kock, 1992).
micronekton community as a low diversity oceanic
faunal assemblage, compared to subtropical and
4.2. Community structure
tropical oceans, where species numbers are four to
five times higher (Hopkins and Lancraft, 1984;
Lancraft et al., 1989). Our data showed a strong separation between
The dominant position of E. antarctica, G. samples taken above and below 450 m depth.
braueri, P. bolini in the Antarctic micronekton These finding could be well correlated with the
assemblage have also been shown for the Scotia presence of warmer and denser CDW at and below
Sea and slope waters off South Georgia (Piat- this depth. Samples from the CDW showed a
kowski et al., 1994). These species are assumed to higher species diversity based on Shannon’s
be the most common myctophids of the Southern diversity and Taxonomic diversity (D). In contrast
the Taxonomic distinctness (DÃ ) showed a high
Ocean (Hulley, 1981). The high densities of E.
antarctica over the slope of King George Island degree of similarity between the deeper and lower
confirmed the importance of this species in the mesopelagic fish community, due to the dom-
pelagic ecosystem south of the Antartic Polar inance of the family Myctophidae and the high
ARTICLE IN PRESS
C. Pusch et al. / Deep-Sea Research I 51 (2004) 1685–1708 1703
number of congeneric species in the lower meso- et al., 1998). In our study, no influence of the shelf-
pelagic assemblage. break on the mesopelagic fish assemblage was
The higher density of mesopelagic fish samples discernable. This could be due to the relatively
taken in the CDW compared to ASW are in small horizontal extent of our study site. There-
accordance with the results of Piatkowski et al. fore, further studies of community structure
(1994), who found a high similarity in species should cover a wider geographical area to fully
composition in lower mesopelagic depths encompass hydrographic features like frontal
(400–1000 m) sampled at two different stations in zones.
the Scotia Sea, while samples taken in the upper
4.3. Feeding ecology
depth horizons were clearly separated geographi-
cally. They found the highest biomass and
4.3.1. Food composition
abundance in the upper layers of the CDW
(400–800 m). Our data showed that Antarctic myctophids,
The analysis of the community structure of like myctophids in other oceans, are primarily
migratory species, like mesopelagic fish, has to zooplanktivores. They feed mainly on copepods
consider the day/night distribution of the species. and euphausiids. The proportions of these main
Because of our restricted sampling program (no prey categories differed in the diets of the species
examined. Large myctophids (460 mm SL), E.
samples above 273 m depth, only four night
antarctica, G. braueri and G. nicholsi fed mainly on
samples) conclusions about the vertical migration
euphausiids, while the smaller-sized P. bolini
patterns of the mesopelagic fish can be drawn only
from data in the literature (Lancraft et al., 1989; (53 mm maximum SL) preyed mainly on cope-
Piatkowski et al., 1994; Duhamel et al., 2000). pods. These results suggest niche separation by the
The results of the multivariate statistical analy- utilization of different prey taxa.
sis of the mesopelagic fish data can be best More than 90% of the krill biomass is concen-
explained by diel migratory behaviour of the trated in the upper 100 m of the water column in
species. The most characteristic species, E. antarc- the South Shetland Island region (Siegel, 1985).
tica and G. braueri, in the group of deeper stations The diet analysis indicated that krill was the main
prey item of E. antarctica and G. nicholsi. It can be
(as defined by SIMPER analysis) perform exten-
assumed that E. antarctica and G. nicholsi were
sive vertical migrations (Torres and Somero, 1988;
Lancraft et al., 1989; Piatkowski et al., 1994). sampled below their main feeding depth, because
During daylight, they stay in the core of the CDW the fishes examined in our study were taken in
(400–800 m) and migrate to the upper 200 m at depths greater than 273 m. This assumption was
supported by the low stomach filling in E.
night. A third species characteristic of deeper
stations was P. gracilis, a species whose vertical antarctica (57.9% empty stomachs) and the
distribution is probably confined to mesopelagic advanced digestion stage of krill found in the
stomachs of G. nicholsi (46.3% digestion stage 3).
depths. The most characteristic species of shallow
samples, P. bolini, also migrates vertically, but In contrast, the proportion of heavily digested
Thysanoessa macrura, a species characterized by a
over a more restricted range as described by
Piatkowski (1989) and Lancraft et al. (1989). wider vertical distribution range (Lancraft et al.,
Therefore, a substantial part of the population 1989), was much smaller (9.1% digestion stage 3).
The high importance of M. gerlachei in the diet of
remains in the ASW (0–400 m depth) during
P. bolini in our study re-confirmed the results of
daylight.
Duhamel et al. (2000) showed a clear change in Ascenio and Moreno (1984). The vertical distribu-
tion maximum of M. gerlachei is in the upper and
the species composition of mesopelagic fish in the
vicinity of the Polar Front at Kerguelen Island. A lower mesopelagic depths (Huntley and Escritor,
frontal system has also been described for the 1992). It can be assumed that this copepod species
shelf-break zone north of King George Island that was available over the whole vertical distribution
range of P. bolini. In contrast to these findings,
results in increased krill densities in this area (Ichii
ARTICLE IN PRESS
C. Pusch et al. / Deep-Sea Research I 51 (2004) 1685–1708
1704
Gaskett et al. (2001) described P. bolini as an E. antarctica taken in samples from November
to January (Rowedder, 1979a) and P. bolini
euphausiid feeder, as this prey category made up
more than the half of the food wet weight in their sampled in February (Ascenio and Moreno,
study that was undertaken close to the subantarc- 1984) have been shown to prey heavily on
tic Maquarie Island. Such differences support the development stages of krill. The breeding season
findings of Pakhomov et al. (1996), namely that of krill fluctuates in the Antarctic Peninsula region
the food composition of myctophids varies by from the beginning of December until the end of
locality and probably by season. The same March (Siegel et al., 1997). During our study, at
conclusion is true for the role of krill in the diets the beginning of December, most female krill
of E. antarctica and G. nicholsi. The importance of individuals were in a pre-spawning or early
adult krill in the diet of both species has been spawning condition (Siegel et al., 1998a). There-
described previously for several areas of the fore, it can be assumed that no krill development
Southern Ocean (Rembiszewski et al., 1978; stages were available in the zooplankton commu-
Rowedder, 1979a; Lubimova et al., 1983; Kozlov nity at the time of our study. In summary, our data
and Tarverdiyeva, 1989). However, Pakhomov et indicate that krill was an important prey item for
al. (1996) found no krill in the diet of E. antarctica E. antarctica and G. nicholsi over the slope of King
and only small quantities in the diet of G. nicholsi George Island. It can be expected that the dietary
sampled in the Prydz Bay, South Georgia and importance of krill for myctophids will be en-
Lazarev Sea (Indian Sector) regions. Therefore, hanced during the course of the austral summer,
these authors suggest that krill only plays a when krill abundance increases and their smaller
significant role in the diet of myctophids only development stages become available in the
during certain time periods and within specific zooplankton (Siegel et al., 1997; Ichii et al.,
1998). Our data also show that P. gracilis is an
areas. An additional reason for the minor im-
portance of krill in the diet of E. antarctica and G. important predator of mesopelagic fish over the
nicholsi in their study could be result of size- slope of King George Island, so confirming the
dependent feeding. Numerous studies have shown results of Rembiszewski et al. (1978).
that larger individuals of a population are able to
4.3.2. Feeding chronology
consume larger prey items, which is often demon-
Except for G. braueri, all myctophids in the
strated by an ontogenetic switch from copepods to
euphausiids as the main prey category (Young and present study showed diurnal feeding patterns
Blaber, 1986; Williams et al., 2001). The size- characterized by increased food intake during
ranges of E. antarctica (25–85 mm SL) and G. night. Nevertheless the day-night differences in
nicholsi (69–139 mm SL) examined by Pakhomov feeding activity were not significant for any
et al. (1996) were smaller than their size-ranges species. Pakhomov et al. (1996) suggested an
in our study, namely 57–113 mm SL and active nocturnal feeding period of 8–10 h for
123–172 mm, respectively. Nevertheless, the Antarctic myctophids. Migratory myctophids,
difference in diet composition cannot be explained especially in nutrient poor tropical and subtropical
entirely by predator length. In the present waters, are reported to show a pronounced
study the diet of the smaller-sized E. antarctica chronology of feeding activity, by feeding in the
(57–79 mm) and G. nicholsi (123–144 mm), that is food rich epipelagic layer at night (Clarke, 1978;
size ranges represented in Pakhomov’s samples, Kinzer and Schulz, 1985; Gartner et al., 1997). In
have a similar food composition to larger contrast, in more productive areas, a number of
migratory myctophids such as Diaphus theta,
individuals and is dominated by euphausiids.
Therefore scarcity of E. superba in the diet of Stenobrachius leucopsarus, Tarletonbeania crenu-
E. antarctica and G. nicholsi in Pakhomov’s laris (Tyler and Pearcy, 1975), Benthosema glaciale
(Kinzer, 1977), Diaphus danae, Lampanyctus hec-
study may be due more to the effect of food
toris (Young and Blaber, 1986) and D. theta
availability than the effect of size-dependent
feeding behaviour. (Moku et al., 2000) tend to feed continuously. It
ARTICLE IN PRESS
C. Pusch et al. / Deep-Sea Research I 51 (2004) 1685–1708 1705
should be noted that the Southern Ocean ecosys- examined were empty, Table 5). Because of this
tem is characterized by a seasonal high productiv- fact, our calculated daily rations could be under-
ity near the pack ice and in the shelf regions like estimated. However, we suggest that our estimates
the slope region of King George Island (Hempel, are realistic, because we have re-calculated prey
1985; Ichii et al., 1998). dry weights by length-weight regressions of un-
In our study, G. nicholsi showed an increased digested prey items. The use of re-calculated
feeding activity in the time period 2100–2400 h, weight is expected to bias the SCI values in a
while in P. bolini SCI was above the average from positive direction (Hyslop, 1980). It is for this
2000 to 0300 h. However, neither species ceased reason that we felt justified in applying the daily
feeding during the day, and both showed a second rations of myctophids in the calculation of their
feeding activity peak in the time interval predation impact on the krill stock in the study
0800–0900 h. A daily feeding period in P. bolini area. Most estimates of the predation impact by
appears to be possible due to the relatively shallow mesopelagic fish on the zooplankton stock have
daytime depth they occupy (273–500 m). This been performed in tropical and subtropical
range overlaps with the vertical distribution range waters. Despite differences in regions and
of its major prey item, M. gerlachei (Hernandez- methods of estimation used, all these studies
Leon et al., 2001). The SCI of E. antarctica indicate that mesopelagic fish are able to remove
increased during the hours of darkness and in two to three times the standing stock of herbivor-
the early morning. In contrast during summer and ous zooplankton (Dalpadado and Gjoesaeter,
autumn at South Georgia, South Sandwich and 1988; Hopkins and Gartner, 1992; Gartner et al.,
South Orkney Island Rowedder (1979a) observed 1997).
the highest feeding activity of this species in the There has been only one previous estimate of
afternoon (1400 h) and second peak in the morning zooplankton consumption by Antarctic mesopela-
(0600 h). The different feeding patterns of E. gic fish. Naumov (1985) as cited in Pakhomov et
antarctica observed by Rowedder suggest that al. (1996) on the basis of an assumed mesopelagic
the feeding chronology of mesopelagic fish can fish stock of 275 million tonnes in the Southern
show a high regional and seasonal variability. Ocean, estimated an annual consumption of 1085
million tonnes of meso- and macro zooplankton
4.3.3. Daily ration and predation impact on the krill (excluding E. superba). This equates to an esti-
stock mated predation impact equivalent to 40% of the
Our estimates of daily rations (0.7–3.3% of dry annual secondary production.
body weight, assuming a 24 h feeding period) were We have calculated the annual removal of the
krill stock in the Southern Ocean region by E.
in good agreement with the estimates of Pakho-
antarctica and G. nicholsi, which are thought to be
mov et al. (1996) for Antarctic myctophids. The
latter ranged from 0.2% to 4.4% of dry body two of the most abundant myctophids in the
weight. Southern Ocean, to be 11.1–26.7%. These results
The estimates of daily ration given by Rowedder confirm the important position of mesopelagic fish
(1979a) for E. antarctica (5% of dry body weight) in the Antarctic oceanic system, and give support
and by Gerasimova (1990) for E. carlsbergi to the suggestion of Lancraft et al. (1989) that they
(3.7–5.6 wet body weight) were higher than in are the most prevalent predators on krill in the
our results. While feeding activity in migratory Antarctic ecosystem. However, the figures given in
mesopelagic fish has been observed to reach the present study should be treated with caution
maximum values in the food-rich epipelagic layers since they represent a ‘‘snapshot’’ of the mid-water
(Gartner et al., 1997), the calculations of daily fish community over the slope of King George
rations in our study were based on individuals Island. Because the feeding impact of mesopelagic
sampled from mesopelagic depths. E. antarctica fish on the krill stock is largely dependent on
and G. braueri, in particular, showed a low feeding locality and season, future studies should be
activity (57.9% and 68.7% of all stomachs carried out at different seasons of the year to get
ARTICLE IN PRESS
C. Pusch et al. / Deep-Sea Research I 51 (2004) 1685–1708
1706
a more realistic picture of the Antarctic mesope- Eggers, D.M., 1977. Factors in interpreting data obtained by
diel sampling of fish stomachs. Journal of the Fisheries
lagic food web.
Research Board of Canada 34, 290–294.
Field, J.G., Clarke, K.R., Warwick, R.M., 1982. A practical
strategy for analysing multispecies distribution patterns.
Acknowledgements Marine Ecology-Progress Series 8, 37–52.
Gartner Jr., J.V., Crabtree, R.E., Sulak, K.J., 1997. Feeding at
Special thanks to Dr. Volker Siegel (Sea Fisheries depth. In: Randall, D.J., Farrel, A.P. (Eds.), Deep-Sea
Fishes. Academic Press, San Diego, pp. 1–388.
Research Institute) for his great assistance in the
Gartner Jr., J.V., Conley, W.J., Hopkins, T.L., 1988. Escape-
identification of zooplankton and his constructive
ment by fishes from midwater trawls: a case study using
comments on an earlier version of this manuscript. lanternfishes (Pisces: Myctophidae). Fishery Bulletin 87,
We also thank Captain Pahl and the crew of the RV 213–222.
‘Polarstern’ for outstanding work at sea. Gaskett, A.C., Bulman, C.M., He, X., Goldsworthym, S.D.,
2001. Diet composition and guild structure of mesopelagic
and bathypelagic fishes near Macquarie Island, Australia.
References NewZealand Journal of Marine Freshwater Research 35,
469–476.
Ascenio, G., Moreno, C.A., 1984. Dieta y selectividad George, E.L., Hadley, W.F., 1979. Food and habitat partiton-
alimentaria de Protomyctophum bolini Fraser-Brunner ing between Rock Bass (Ambloplites rupestris) and small-
(Pisces; Myctophidae) en el paso Drake (Antarctica). Serie mouth Bass (Micropterus dolomieui) young of the year.
cientifica. Instituto Antartico chileno. 31, 85–91. Transactions of the American Fisheries Society 108,
Barlow, K.E., Boyd, I.L., Croxall, J.P., Reid, K., Staniland, 253–261.
I.J., Brierley, A.S., 2002. Are penguins and seals in Gerasimova, O.V., 1990. Feeding and food intake of Electrona
competition for Antarctic krill at South Georgia? Marine carlsbergi (Taning, 1932) Myctophidae. Selenium Science
Biology 140, 205–213. Paper CCAMLR 7, 411–416.
Bulman, C.M., He, X., Koslow, J.A., 2002. Trophic ecology of Gon, O., Heemstra, P.C., 1990. Fishes of the Southern Ocean.
the mid-slope demersal fish community off southern J.L.B. Smith Institute of Ichthyology, Grahamstone
Tasmania, Australia. Marine Freshwater Research 53, (462pp).
59–72. Greely, T.M., Gartner Jr., J.V., Torres, J.J., 1999. Age and
Cherel, Y., Guinet, C., Tremblay, Y., 1997. Fish prey of growth of Electrona antarctica (Pisces: Myctophidae), the
Antarctic fur seals Arctocephalus gazella at Ile de Croy, dominant mesopelagic fish of the Southern Ocean. Marine
Kerguelen. Polar Biology 17, 87–90.
Biology 133, 145–158.
Clarke, K.R., Ainsworth, M., 1993. A method of linking
Groeger, J., Piatkowski, U., Heinemann, H., 2000. Beak length
multivariate community structure to environmental vari-
analysis of the Southern Ocean squid Psychroteuthis
ables. Marine Ecology-Progress Series 92, 205–219.
glacialis (Cephalopoda: Psychroteuthidae) and its use for
Clarke, K.R., Warwick, R.M., 1994. Change in Marine
size and biomass estimation. Polar Biology 23, 70–74.
Communities: An Approach to Statistical Analysis and
Guinet, C., Cherel, Y., Ridoux, V., Jouventin, P., 1996.
Interpretation. Natural Environmental Research Council,
Consumption of marine resources by seabirds and seals in
Plymouth (p. 144).
Crozet and Kerguelen waters: changes in relation to
Clarke, K.R., Warwick, R.M., 2001. Primer-E (5) Computer
consumer biomass 1962–1985. Antarctic Science 8, 23–30.
Program. Natural Environmental Reasearch Council,
Hempel, G., 1985. Antarctic marine food webs. In: Siegfried,
Plymouth.
W.R., Condy, P.R., Laws, R.M. (Eds.), Antarctic Nutrient
Clarke, T.A., 1978. Diel feeding patterns of 16 species of
Cycles. Springer, Berlin, pp. 266–276.
mesopelagic fishes from Hawaiian waters. Fishery Bulletin
Hernandez-Leon, S., Portillo-Hahnefeld, A., Almeida, C.,
76, 495–513.
Becognee, P., Moreno, I., 2001. Diel feeding behaviour of
Dalpadado, P., Gjoesaeter, J., 1988. Feeding ecology of the
krill in the Gerlache Strait, Antarctica. Marine Ecology-
lanternfish Benthosema pterotum from the Indian Ocean.
Progress Series 223, 235–242.
Marine Biology 99, 555–567.
Hoddell, R.J., Crossley, A.C., Williams, R., Hosie, G.W., 2000.
Duhamel, G., 1998. The pelagic fish community of the Polar
The distribution of Antarctic pelagic fish and larvae
frontal zone of the Kerguelen Islands. In: Di Prisco, G.,
(CCAMLR division 58.4.1). Deep-Sea Research Part II
Pisano, E., Clarke, A. (Eds.), Fishes of Antarctica: A
47, 2519–2541.
Biological Overview. Springer-Verlag Italia, Milano (Italy),
Hofmann, E.E., Klinck, J.M., Lascara, C.M., Smith, D.A.,
pp. 63–74.
1996. Water mass distribution and circulation west of the
Duhamel, G., Koubbi, P., Ravier, C., 2000. Day and night
Antarctic Peninsula and including Bransfield Strait. Ant-
mesopelagic fish assemblage off the Kerguelen Islands
(Southern Ocean). Polar Biology 23, 106–112. arctic Research Series 70, 61–80.
ARTICLE IN PRESS
C. Pusch et al. / Deep-Sea Research I 51 (2004) 1685–1708 1707
Lubimova, T.G., Shust, K.V., Troyanovskij, F.M., Semenov,
Hopkins, T.L., Gartner Jr., J.V., 1992. Resource partitioning
A.B., 1983. The ecology of mass species of myctophids from
and predation impact of a low latitude myctophid commu-
Atlantic Antarctica. Antarktika 22, 99–106.
nity. Marine Biology 114, 185–197.
McGinnis, R.F., 1982. Biogeography of laternfishes (Mycto-
Hopkins, T.L., Lancraft, T.M., 1984. The composition and
phidae) south of 301. Antarctic Research Series 35, 1–110.
standing stock of mesopelagic micronekton at 271N 861W in
Mizdalski, E., 1988. Weight and length data of zooplankton in
the eastern Gulf of Mexico. Contributions to Marine
the Weddell Sea in austral spring 1986 (ANTV/3). Berichte
Science 27, 143–158.
zur Polarforschung 55, 1–72.
Hulley, P.A., 1981. Results of the research cruises of FRV
Moku, M., Kawaguchi, K., Watanabe, H., Ohno, A., 2000.
‘‘Walther Herwig’’ to South America LVIII. Family
Feeding habits of three dominant myctophid fishes, Diaphus
Myctophidae (Osteichthyes, Myctophiformes). Archiv fur ¨
theta, Stenobrachius leucopsarus and S. nannochir, in the
Fischereiwissenschaft 31, 1–300.
subarctic and transitional waters of the western North
Hulley, P.A., 1998. Preliminary investigations on the evolution
Pacific. Marine Ecology-Progress Series 207, 129–140.
of the tribe Electronini (Myctophiformes, Myctophidae). In:
Nakamura, I., 1990. Gempylidae. In: Gon, O., Heemstra, P.C.
Di Prisco, G., Pisano, E., Clarke, A. (Eds.), Fishes of
(Eds.), Fishes of the Southern Ocean. JLB Smith Institute of
Antarctica: a biological overview. Springer-Verlag Italia,
Ichthyology, Grahamstone, South Africa, pp. 402–403.
Milano, pp. 75–85.
Naumov, A.G., 1985. System analysis of the structure of the
Hulley, P.A., Kock, K.H., Pusch, C., White, M.G., 1998.
Antarctic waters community. In: Vinogradov, M.E., Flint,
Mesopelagic fishes. Berichte zur Polarforschung 274, 17–22.
Huntley, M.E., Escritor, F., 1992. Ecology of Metridia gerlachei M.V. (Eds.), Biological Basis of the Commercial Utilization
of the Open Regions of the Ocean. Nauka Press, Moscow,
Giesbrecht in the western Bransfield Strait, Antarctica.
pp. 57–80 (in Russian).
Deep-Sea Research A 39, 1027–1055.
Pakhomov, E.A., Perissinotto, R., McQuaid, C.D., 1996. Prey
Hyslop, E.J., 1980. Stomach contents analysis-a review of
composition and daily rations of myctophid fishes in the
methods and their application. Journal of Fish Biology 17,
Southern Ocean. Marine Ecology-Progress Series 134, 1–14.
411–429.
Pearcy, W.G., Lorz, H.V., Peterson, W., 1979. Comparision of
Ichii, T., Naganobu, M., Ogishima, T., 1996. Competition
the feeding habits of migratory and non-migratory Steno-
between the krill fishery and penguins in the South Shetland
brachius leucopsarus (Myctophydae). Marine Biology 51,
Islands. Polar Biology 16, 63–70.
1–8.
Ichii, T., Katayama, K., Obitsu, N., Ishii, H., Naganobu, M.,
1998. Occurrence of Antarctic krill (Euphausia superba) Phillips, K.L., Jackson, G.D., Nichols, P.D., 2001. Predation
on myctophids by the squid Moroteuthis ingens around
concentrations in the vicinity of the South Shetland Islands:
Macquarie and Heard Islands: stomach contents and fatty
relationship to environmental parameters. Deep-Sea Re-
acid analysis. Marine Ecology-Progress Series 215, 179–189.
search I 45, 1235–1262.
Piatkowski, U., 1989. Macroplankton communities in Antarctic
Kinzer, J., 1977. Observations on feeding habits of the
mesopelagic fish Benthosema glaciale (Myctophidae) off surface waters: spatial changes related to hydrography.
Marine Ecology-Progress Series 55, 251–259.
NW Africa. In: Anderson, N.R., Zahuranec, B.J. (Eds.),
Piatkowski, U., Rodhouse, P.G., White, M.G., Bone, D.G.,
Oceanic Sound Scattering Prediction. Plenum Press, New
Symon, C., 1994. Nekton community of the Scotia Sea as
York, pp. 381–392.
sampled by the RMT 25 during austral summer. Marine
Kinzer, J., Schulz, K., 1985. Vertical distribution and feeding
Ecology-Progress Series 112, 13–28.
patterns of midwater fish in the central equatorial Atlantic.
Pielou, E.C., 1975. Ecological Diversity. Wiley, New York
I. Myctophidae. Marine Biology 85, 313–322.
(165pp).
Kock, K.-H., 1992. Antarctic Fish and Fisheries. Cambridge
Rembiszewski, J.M., Krzeptowski, M., Linkowski, T.P., 1978.
University Press, Cambridge (359pp).
Fishes (Pisces) as by-catch in fisheries of krill Euphausia
Kock, K.-H., Dollefeld, L., Hulley, P.A., Jahrig, H.-P., Petzel,
¨ ¨
superba Dana (Euphausiacea, Crustacea). Polskie archiwum
W., Pusch, C., Siegel, V., White, M.G., 1998. The
hydrobiologii 25, 677–695.
composition of the fish fauna on the northern slope of
Rodhouse, P.G., White, M.G., Jones, M.R.R., 1992. Trophic
King George Island. Berichte zur Polarforschung 274,
relations of the cephalopod Martialia hyadesi (Teuthoidea:
14–17.
Ommastrephidae) at the Antarctic Polar Front, Scotia Sea.
Kozlov, A.N., Tarverdiyeva, M.I., 1989. Feeding of different
Marine Biology 114, 415–421.
species of Myctophidae in different parts of the Southern
Rowedder, U., 1979a. Feeding ecology of the myctophid
Ocean. Journal of Ichthyology 29, 160–167.
Electrona antarctica (Guenther, 1878) (Teleostei). Meeres-
Lancraft, T.M., Torres, J.J., Hopkins, T.L., 1989. Micronekton
forschung 27, 252–263.
and Macrozooplankton in the open waters near Antarctic
Rowedder, U., 1979b. Some aspects of the biology of Electrona
Ice Edge Zones (AMERIEZ). Polar Biology 9, 225–233.
antarctica (Guenther, 1878) (Family Myctophidae). Meer-
Lancraft, T.M., Hopkins, T.L., Torres, J.J., Donnelly, J., 1991.
esforschung 27, 244–251.
Oceanic micronecton/macrozooplanktonic community
Schlitzer, R., 2003. Ocean data view. http://www.awi-bremer-
structure and feeding in ice covered Antarctic waters during
haven.de/GEO/ODV.
the winter (AMERIEZ 1988). Polar Biology 11, 157–167.
ARTICLE IN PRESS
C. Pusch et al. / Deep-Sea Research I 51 (2004) 1685–1708
1708
Shannon, C.E., Weaver, W., 1949. The Mathematical Theory of Torres, J.J., Somero, G.N., 1988. Vertical distribution and
Communication. University Illinois Press, Chicago (p. 111). metabolism in Antarctic mesopelagic fishes. Comparative
Siegel, V., 1985. The distribution pattern of krill, Euphausia Biochemistry and Physiology 90B, 521–528.
superba, west of the Antarctic Peninsula in February 1982. Tyler Jr., H.R., Pearcy, W.G., 1975. The feeding habits of three
Meeresforschung 30, 292–305. species of laternfishes (Family Myctophidae) off Oregon,
Siegel, V., de la Mare, W.K., Loeb, V., 1997. Long-term USA. Marine Biology 32, 9–11.
monitoring of krill recruitment and abundance indices in the Warwick, R.M., Clarke, K.R., 1995. New ‘biodiversity’
Elephant Island area (Antarctic Peninsula). CCAMLR measures reveal a decrease in taxonomic distinctness with
Science 4, 19–35. increasing stress. Marine Ecology-Progress Series 129,
Siegel, V., Kock, K.H., Petzel, W., Hulley, P.A., Pusch, C.,
301–305.
Dollefeld, L., White, M.G., 1998a. Investigation on the krill
¨
Williams, A., Koslov, J.A., Terauds, A., Haskard, K., 2001.
stock and the epipelagic zooplankton community in the
Feeding ecology of five fishes from the mid-slope micro-
Elephant Island region. Berichte zur Polarforschung 274,
nekton community off southern Tasmania, Australia.
50–56.
Marine Biology 139, 1177–1192.
Siegel, V., Loeb, V., Groeger, J., 1998b. Krill (Euphausia
Williams, R., 1985. Trophic relationships between pelagic
superba) density, proportional and absolute recruitment and
fish and euphausiids in Antarctic waters. In: Siegfried,
biomass in the Elephant Island region (Antarctic Peninsula)
W.R., Condy, P.R., Laws, R.M. (Eds.), Antarctic
during the period 1977–1997. Polar Biology 19, 393–398.
Nutrient Cycles and Food Webs. Springer, Berlin,
Stein, M., Heywood, R.B., 1994. Antarctic environment-
pp. 452–459.
physical oceanography: the Antarctic Peninsula and South-
Young, J.W., Blaber, S.J.M., 1986. Feeding ecology of three
west Atlantic region of the Southern Ocean. In: El-Sayed,
species of midwater fishes associated with the continental
S.Z. (Ed.), Southern Ocean Ecology—The Biomass Per-
slope of eastern Tasmania, Australia. Marine Biology 93,
spective. Cambridge University Press, Cambridge, pp.
11–24. 147–156.