Trophodynamics of the Hyperiid Amphipod Themisto Gaudichaudi in the South Georgia Region During Late Austral Summer

l MARINE ECOLOGY PROGRESS SERIES Vol. 134: 91-100.1996 ' Published April 25 Mar Ecol Prog Ser

Trophodynamics of the hyperiid amphipod Themisto gaudichaudi in the South Georgia region during late austral summer

E. A. Pakhomov*, R. Perissinotto

Southern Ocean Group, Department of Zoology & Entomology. Rhodes University. PO Box 94, Grahamstown 6140. South Africa

ABSTRACT. The feeding dynamics and predation impact of the hyperiid amphipod Themisto qau- dichaudi on mesozooplankton were studied in the vicinity of South Georgia, Southern Ocean, during austral summer 1994. Data show that 7, gaudichaudi is a visual opportunistic predator, consuming pri- marily the most abundant species of copepods, euphausilds and pteropods. In situ estimated daily rations were equivalent to 6.3 % of body dry weight, and similar to the value of 7.1 % of body dry weight derived using an energy budget approach which takes into account the total daily energy requirements. In vitro estimates produced daily rations higher than these, ranglng from 8.5 to 21.8% of body dry weight, increasing with prey density. The predation impact of T. gaudichaudiadults, averaged over a 0 to 200 or 0 to 100 m layer, never exceeded 2.1 of mesozooplankton standing stock per day but accounted for up to 70% of the daily secondary production This suggests that in the vicinity of South Georgia T. qaudichaudiadults are able to control the local mesozooplankton community and may contribute significantly to the downward flux of biogenic carbon.

KEYWORDS: Antarctica South Georgia - Amphipods . Themistogaudichaudi - Feeding. Daily ration. Predation impact

INTRODUCTION with the northern hemisphere species may be applica- ble to the ecologically similar species of the Southern The hyperiid Themisto gaudichaudi is one of the Ocean, independent estimates of daily ration and pre- most ab.undant pelagic amphipods of the Southern dation impact of T gaudichaudi in Antarctic waters are Ocean, occurring in large numbers within the sub- still not available. antarctic and low Antarctic zones (Hardy & Gunther Amphipods of the genus Themisto are generally 1935, Kane 1966, Pakhomov & McQuaid 1996). Previ- regarded as obligate carnivorous species (Kane 1967, ously regarded as a species with a bipolar distribution, Sheader & Evans 1975, Hopkins 1985) and are also in 1986 the Antarctic and the Arctic 'forms' of T. gau- known to form parasltic/parasitoid associations with dichaudi were separated into 2 specles by Schneppen- gelatinous zooplankton like salps and medusae (Ma- heim & Weigmann-Haass (1986) using morphological din & Harbison 1977, Lava1 1980). There are however and electrophoretic evidence. Several feeding studies some indications that they may consume phytoplank- with T gaudichaudi (now T. compressa) and with other ton cells as well, especially in the early stages of devel- members of the genus Themisto have been carried out opment (Siegfried 1965, Nemoto & Yoo 1970, Hopkins in the northern hemisphere (Sheadel- & Evans 1975, 1985, Sugisaki et al. 1991). T. gaudichaudi has a high Yamashita et al. 1984, 1985, Persy 1993). Although it is energy content (Williams & Robins 1979, Strelnikova generally assumed (Ikeda 1989) that results obtained 1989, Torres et al. 199413) and is one of the main food sources for many predators including fish, squid, birds and whales (Nemoto & Yoo 1970, Permitin & Tarver- dieva 1972, Chechun 1984, Rodhouse et al. 1992, Bost

0 Inter-Research 1996 Resale offull article not permitted 92 Mar Ecol Prog Ser 134: 91-100, 1996

et al. 1994, Kock et al. 1994).Thus, T gaudichaudi con- area of 0.5 m2 and a mesh size of 0.3 mm. Both nets stitutes an important ecological link between small were fitted with a Universal Underwater Unit (Robert- zooplankton and top consumers and, in certain areas, son et al. 1981). The volume filtered by the Bongo net may effectively control the mesozooplankton standing was calculated using electronic flowmeter (General stock (Gibbons et al. 1992) and affect the population Oceanics, Miami, FL, USA) data, while for the WP-2 d.ynamics of predators such as penguins (Bost et al. this was determined by multiplying the mouth area by 1994) and fish (Kock et al. 1994). T. gaudlchaudi may the distance sampled. Bongo nets were towed also contribute substantially to the downward carbon obliquely between 0 and 200 m while 0 to 100 m verti- flux in the subantarctic zone due to its high abun- cal hauls were made with the WP-2 net. In addition, dances, marked die1 vertical migrations (Kane 1966, collections with a Neuston net (mesh size 0.8 mm) were Everson & Ward 1980) and production of fast-sinking obtained on 2 occasions by towing at the surface for faecal pellets (Fortier et al. 1994). -30 min during darkness. Specimens of T. gaudichaudi Themisto gaudichaudi is the d0minan.t pelagic am- for gut content analysis were also taken from krill sam- phipod in the waters surrounding South Georgia, ples collected with the Polish Krill Trawl 1641 (mesh where it often occurs in dense swarms (Hardy & Gun- size of 7 mm). Trawls were towed obliquely between ther 1935, Atkinson & Peck 1988). Nevertheless, feed- the surface and -100 m depth and, on a few occasions, ing studies on this species have never been under- as close as possible to the shelf bottom. All samples taken in the region. The main aim of this paper is, were preserved in 4 to 6% buffered formalin and therefore, to provide some estimates of dally ration and examined in the laboratory. predation impact of T. gaudichaudi on the mesozoo- Length-weight relationships. Three to 25 preserved plankton community of South Georgia. individuals Themisto gaudichaudi were taken from each sample for length-weight analyses. Three different body length measurements were undertaken with MATERIALS AND METHODS an accuracy of -0.01 mm: from the anterior part of the head, excluding the antenna, to the end of the uro- Sampling strategy. Specimens of Themisto gaudi- pods (L,);to the tip of the telson (L,); and to the end chaudi were obtained from general zooplankton sam- of the thorax (L3).After measurements, each animal ples and from tows of krill trawls made during voyage was dried on filter paper and its wet weight deter- 119 of the RV 'Africana'in February-March 1994 near mined with an accuracy of -0.1 mg. Dry weight was South Georgia (Fig. 1). During the survey, zooplankton estimated by weighing hyperiids oven-dried at 60°C was sampled using Bongo and WP-2 nets with a mouth for 36 h.

15875 (T) 15891 (B) a l5880 (B) • 15905 (T) 15873(B) 15906 (W) l5898 (T) 15907 (W) 15909 (N) 15860 (T) 15910 (H') 15912 (W) - 54' - 15889 (B) 4 a 15914 (T) 15884 (B) m m@ 15904 (T) 15859 (B)

15855 (T) - -ss 15863 (B)

I I I I a l 39" 38" 37" 36" 35" LONGITUDE(W)

Fig. 1. Position of sampling stations in the vicinity of South Georgia during February-March 1994. Samples were collected using either a Bongo net (B),a WP-2 net (W),a Neuston net (N) or a Polish Knll Trawl (T) Pakhomov & Perissinotto: Trophodynamics of Themisto gaudlchaudi

Gut content analysis. From 4 to 25 (depending on P = 0.43 X R. The food intake is: C = P+R/U, where U is total sample size) preserved Themisto gaudichaudi the assimilation coefficient (Vinberg 1956). Assuming from each tow were examined for gut contents. Hyper- U-0.8 (Vinogradov & Shushkina 1987), a value for Cof iids were first measured from the anterior part of the -1.79 X R is derived. Under natural conditions, R is head to the end of the thorax. After their gut content equal to -2 X Rbw,where R,, represents the basic meta- was removed, animals were oven-dried at 60°C for bolic needs during winter. According to estimates 36 h. Prey items were isolated from stomachs and, obtained by Torres et al. (1994a, b), 1 g wet weight of where possible, identified to species level. One to 5 gut T gaudichaudi contains 687 cal and R,, = 0.133 p1 0, contents, depending on stomach fullness, were then mg-' wet weight h-', or (using average wet weight of T. collected onto a preweighed GF/F filter and dried at gaudichaudi: 0.073 g) 0.233 m1 O2 ind-l d-' Assuming 60°C for 24 h. All results from the stomach content 1 m1 O2 = 4.86 cal = 7.07 mg wet weight of T. gau- analysis were expressed as frequency of occurrence dichaudi, R can be estimated by doubling the basic (%) of each item in stomachs with food. The index of metabolic requirements: R = 2 X R,,,, = 2 X 0.233 X stomach fullness, expressed as percentage of body 0.00707 X 100/0.073 = 4.51 % of body wet weight. weight, was calculated by dividing the dry weight of Predation experiments. Three separate predation the stomach contents by the dry weight of the hyperiid experiments were conducted during the cruise period. body. Hyperiids were obtained from 0 to 200 m oblique Gut pigment contents and evacuation rate experi- Bongo tows and only active and undamaged individu- ments. From each tow, modal-sized Themisto gau- als from the 3 l open cod-end were used. Prior to each dichaudi were taken immediately after capture. One to experiment, Themisto gaudichaudi were kept for 24 6 replicate animals were placed in plastic centrifuge to 26 h in 20 1 containers filled with filtered seawater. tubes (one per tube) with 5 m1 of 100% methanol and For each experiment, water containing zooplankton stored at -20°C for 12-24 h (Simard et al. 1985). After previously screened through a 1 mm (Expts 1 and 3) centrifugation (1745 x g), the plgment content of the or 3 mm (Expt 2) mesh was mixed in 20 1 buckets and methanol extract was measured with a Turner 111 flu- then used to fill one 2 1 (experimental) and three 1 1 orometer, before and after acidification (Mackas & (control) volume plastic jars. Three initial (control jars) Bohrer 1976). Pigment contents were expressed in subsamples of 1 1 each were concentrated and pre- terms of total pigments per individual and calculated served in 4% buffered formalin. Experimental ani- according to Strickland & Parsons (1968) as modified mals were then gently transferred to the 2 1jars (4 or 5 by Conover et al. (1986). animals per jar) and incubated for -24 h to integrate To estimate the gut evacuation rate constant (k), any die1 feeding activity which may occur in nature. freshly caught and healthy, modal-sized T. gaudi- Experimental jars were kept on deck at ambient tem- chaudi were taken from the 3 1 cod-end of the Bongo perature and -50% of ambient light intensity, and net and placed into 20 l containers with filtered seawa- agitated constantly by keeping them in a circulating ter. Containers with experimental animals were kept bath. At the end of the experiment, the content of the on deck at ambient temperature. Experiments were experimental jars was concentrated and preserved in run for -8 h and the decline in gut content was moni- 4 "/o buffered formalin. In the laboratory, all zooplank- tored at intervals of 15 to 30 min, for the first 3 h, and of ton were measured with an accuracy of -0.5 mm, 1 to 2 h thereafter. Gut evacuation rate (k, h-') was identified to genus level and counted. After this, the derived from the slope of the regression of the natural dry weights of the zooplankton in the control bottles, logarithm of gut pigments versus time (Dam & Peter- of those remaining in the experimental jars and of T son 1988). gaudichaudi were measured by oven-drying the sam- Estimation of daily ration. Daily ingestion rates of ples at 60°C for 36 h. Themisto gaudichaudi (C,) were estimated from the relation: C,v = Gx 24 X k, where k is the gut evacuation rate constant in h-' and G is the average value of the RESULTS 24 h-integrated (day-cycle) index of stomach fullness in % of body dry weight. Another estimate of daily Length-weight relationships ration was obtained using an energy budget approach, based on estimates of hyperiid metabolic energy Due to a lack in the literature of empirical relation- requirements and Net Conversion Efficiency, K2 = ships between the different types of length measure- P/(P+R), where P is production and R is the metabolic ments generally used (L,, L? and L3), regressions energy demands. K, values for pelagic zooplankton between them were calculated in this study (Table 1). species range from 0.2 to 0.5 and for adults are usually Linear regression models were highly significant (p i equal to -0.3 (Vinogradov & Shushkina 1987). Thus, 0.001) and provided the best fit to the data, explaining 94 Mar Ecol Prog Ser 134: 91-100, 1996

Table 1 Themisto gaudichaudi(N = 69). Regression coeffi- Copepods were the dominant components of the diet cients for the relationships between different length rneasure- of Themisto gaudichaudi at almost all stations, ac- rnents (L,. L,, L,, in mm),wet (W, In mg) and dry (DW, in counting for 10 to 100% (mean 69.2%) of the total prey rng) body weights. All regressions are significant at p < 0.001 occurrence in stomachs (Table 2). Among them, Cala- nus simillimus, Calanoides acutus and Metridia spp. I Type of regression a b l were found most often. The next most abundant prey were euphausiids, mainly furcilia stages of Euphausia spp. and Thysanoessa spp., which accounted for up to 50% of the total food items identified, in those stomachs with food. The pteropod Limacina spp. formed the bulk of T, gaudichaudi diet at least at 2 stations where it constituted up to 50 to 100% of the total food items consumed (Table 2). Ostracods, polychaetes and chae- tognaths were found only occasionally in the stomachs of T. gaudichaudi. The mean index of stomach fullness ranged widely, from 0.26 to 4.54 % of the dry body weight (Table 2). Highest levels coincided with samples in which the lowest proportions of empty stomachs of Themisto >90% of the total variance. The average water content gaudichaudi were recorded. Indices of stomach full- of individual Themisto gaudichaudi was 16.7 * 3.4 % of ness monitored throughout a 24 h day-cycle exhibited the total wet weight. Generally, water content showed highest levels in the early morning hours, just after an inverse linear correlation with hyperiid size sunrise, and during the afternoon before sunset (Table 1).Weight versus length regressions were in the (Fig. 3). Based on the data presented in Fig. 3, the aver- form of a power function and highly significant (p < age index of stomach fullness integrated over a 24 h 0.001), explaining over 85Y0 of total variance (Table 1). day-cycle (G) was 1.96% of body dry weight. Individual levels of gut pigment concentration of Themisto gaudichaudi estimated fluorometrically var- Gut content analyses ied markedly, ranging from 4 to 577 ng (pigm.) ind.-' (Fig. 3). The highest levels of gut pigment contents Themisto gaudichaudi individuals used for gut con- were recorded in the late afternoon (13:OO to 18:OO h tent analysis ranged in size from 6 to 10 mm thorax GMT) (Fig. 3). 1en.gth (L3),with a modal size of 7 to 8 mm (Fig. 2). All data obtained from the gut evacuation experi- Average wet and dry weights were 0.073 and 0.012 g, ments were best fitted by a negative exponential respectively. model of gut pigment content versus time (Fig. 4).The values of gut evacuation rate constant (k) and gut turnover time (l/k) calculated from the model fit were 0.133 h-' and 7.5 h, respectively.

Daily rations

Using the k and G values calculated above, the daily food intake of The.misto gaudichaudi adults was estimated as: C,, = 1.96 X 24 X 0.133 = 6.3% of body dry weight. This is also equivalent to body C as the carbon content of the main food items and of T gaudichaudi adults are very similar (Ikeda & Mitchell 1982, Torres et al. 1994b). The second approach is based on estimates of the 567891011 metabolic requirements of Themisto gaudichaudi and THORAX LENGTH (mm) of its K*. According to these, C, = 1.79 X R = 8.08% of body wet weight. Since the energy spent on reproduc- Fig. 2. Themisto gaudlchaudi.Size-frequency distnbution of tion by crustaceans accounts for -16% of the total daily hypenlds used In the gut content analys~s ration (Tseitlin 1988), C,, would then increase to 9.37 % Table 2. Themisto gaudichaudi. Frequency of occurrence of prey items in stomachs of hyperiids with food present, South Georgia, February-March 1994

Stomach contents Station number 15855 15858 15859 15873 15880 15887 15894 15898 15899 15903 15907 15910 15912 15914

Copepoda Calanus sjm~ll~rnus Calanoides acutus C1ausocalanus spp. Metridia spp. Euchaeta spp. Calanoida (unident.) Ostracoda Conchoecinae Euphausiacea Euphausia fr~gida(fur) Euphausia spp. (fur) Thysanoessa spp. (fur) Euphausiacea (legs) Polychaeta Pelagobia spp. Polychaeta (unident.) Molluscs Limdcina spp. Chaetognatha Eukrohnia spp. Chaetognatha (unident )

Number of stomachs % of empty stomachs Mean index of stomach fullness (X) 0.93 1.86 2.13 3.63 0.26 0.18 0.94 044 1.03 1.70 2.55 4.54 2.35 2.15 96 Mar Ecol Prog Ser 134: 91-100, 1996

m - 1- A The results of the predation experiments are Em- presented in Fig. 5 and Table 3. Themistogau- -P A 0 dichaudiadultspreferentially consumed prey in C 400- C the size range of 1 to 4 mm. However, furcilia ZW and juvenile stages of euphausiids, with total l- 300- Z length of 9 to 13 mm, were also consumed very 8 - A C m- A efficiently (Fig. 5, Expt 2). No feeding selectivity zW E A A f * :i between different genera exhibiting the same p 100- 0 - A A -5 size was observed. Daily rations of T. gau- l- - A--f :t r i dichaudi increased with the increase in prey ~a'i . i 6 i 1 i4 $8-m %* 24 TlME (GMT) density, from 8.5 to 21.8 % of body dry weight in g 12- experimental jars with initial prey density of 354 e A and 1917 ind. 1-', respectively (Table 3, Fig 5). A

E - Predation impact m A *W 6-_ - 3 4 a A To examine the predation impact of Themisto 4- .. Yi A gaudichaudi adults, we used data on mesozoo- 0I fA -A plankton biomass (including copepods, ostra- r. -A * -A f A A E tAA A cods, euphausiids

DISCUSSION

Feeding ecology

The results of the gut content analysis show a close relationship between the structure of the local zooplankton community and the diet composition of Themistogaudichaudi. The most abundant species of TlME (hours) copepods, euphausiids and pteropods collected with Flg. 4 Thernisto gaudichaudi. Gut evacuation rate of adults net tows around South Georgia formed the bulk of the collected in the vicinity of South Georgia during February diet of T ga~ldichaudiadults during austral summer 1994 1994. Previously published data (Siegfried 1965, Pakhomov & Perissinotto: Trophodynamics of Themislo gaudichaudi

250, I shown that phytoplankton may play a substantial role in the feeding budget of young \ Exp. 1 T gaudichaudi (Siegfried 1965, Nemoto & 2 ?""l Initial densitv of food items: 354f33 ind.1.' Yoo 1970, Hopkins 1985).Large amounts of - I 51) I Daily consumplion rate: 8.5%dry body weight ] - phytoplankton pigments were recovered in : loo the guts of T gaudichaudi adults during 3:- this study but it was not possible to show, Z with our technique, whether the phytoplankton cells originated from consumed

-150 - It has been clearly shown in Themisto

400 - japonica that gut pigment concentrations of young hyperiids are significantly higher 3iO- - Exp. 2 27011- Initial density of food items: 1077f95 ind.1" than those of larger specimens (Sugisaki et - Daily consumption rate: 20.3% dry body weight al. 1991). This suggests that the feeding - 250- 200- behaviour of T.japonica switches from her- ;LL: bivory to carnivory during its life-cycle. 150- ' The same has also been proposed for T gaudichaudi in the Benguela Upwelling System (Siegfried 1965). Only adults of T. ; 2- ;c17 '16-17~ gaudichaudi were analyzed in our study 1-2 .?-L 5 7-8 -111-12 13-14 15-16 and no identifiable phytoplankton cells I 410 were recovered from their stomachs. Thus, 121n- the occurrence of sizable levels of pig- ,? ,? Exp. 3 3I0IM)- ments in their guts suggests that the phyto- - Initial densitv of food items: 1917t132 ind.1" plankton material is most likely of sec- 800- Dailv consumption rate: dry body weight U 21.8% 5 W- ondary denvation, e.g. pigments extracted ~ from guts of T gaudichaudi prey. This r 400- z hypothesis is supported by the close covari- 200 - ance between maximum gut pigment lev-

(lh S 7 I 2- I 61, els and total gut contents observed during I-? 4 -6 74 9-10 11-12 13-14 1.5-16 a 24 h day-cycle (Fig. 3). L~NGIH(MM) The gut pigment content data show only HE~OKEINCUHKI'ION [7 A~T~KINCIIBAI'ION 1 clear peak in the diurnal feeding activity of Themistogaudichaudi adults, while the Fig. 5.Themisto gaudichaudi. Size-frequency distribution of the prey items stomach fullness indices exhibit 2 distinct of hyperiids used in the predation experiments peaks. Since pigments in the gut of T.gaudichaudi have a secondary origin, this dis- crepancy is probably the result of the die1 Nemoto & Yoo 1970, Hopkins 1985) and the results of periodicity in the feeding pattern of the hyperiid preys. our predation experiments confirm the hypothesis that The 2 peaks coincided with the periods just after sun- T. gaudichaudi is an opportunistic predator with the rise and before sunset, reflecting the well-known pat- capability of consuming any prey of appropriate size tern of diurnal vertical migration of T. gaudichaudi and taxonomy (Gibbons et al. 1992). It has also been (Hardy & Gunther 1935, Kane 1966, Everson & Ward

Table 3. Themisto gaudichaudi. Results of predation experiments with different prey concentrations

Expt Themisto gaudichaudi Initial prey density (iSD) Final prey density Daily ration no. No. per jar Total DW (mg) (ind. l-l) (mg DW I-') (ind. I-') (mg DW 1-l) (% body DW) Mar Ecol Prog Ser 134: 91-100. 1996

Table 4 Themisto gaudichaudi. Predation impact on mesozooplankton standing stock in the vicinity of South Georgia during Feb- ruary-March 1994. Zooplankton biomass was estimated from oblique Bongo and vertical WP-2 net tows (H. Verheye unpubl. data). The daily ration of adults was assumed to be equal to 6.3 % body dry weight. Secondary production rates were obtained from the P/B (production/biomass) coefficients (-3 % d-') for mesozooplankton available in the literature (Vinogradov & Shushkina 1987)

Stn no. Mesozooplankton Daily 2nd Themisto gaudicha udi Predation impact biomass production Abundance Biomass % % 2nd (g DW m-*) (g DW (ind. m-?) (g DW m-2) biomass production

15859 8.419 0.252 305.0 2.806 2.10 70.2 15863 15.151 0.454 66.5 0.796 0.33 11.0 15873 25.566 0.767 42.6 0.510 0.13 4.2 15880 6.806 0.204 34.4 0.412 0.38 12.7 15884 9.510 0.274 0.8 0.010 <0.01 0.2 15889 7.506 0.225 0 0 0 0 15894 5.196 0.156 13 0 0.104 0.19 6.4 15907 3.655 0.109 47 1 0 564 0.96 32.1 15910 4.325 0.130 78.4 0 940 1.36 45.4 15912 6.987 0.210 15.7 0 188 0.17 5.7

1970). According to this pattern, maximum feeding reported in the literature, indicates that this is very activity would be observed when hyperiids rise similar to the k-values for Euphausia superba, 0.101 to towards the surface to feed and when they sink to 0.424 h-' (Perissinotto & Pakhomov in press) and of T. depth in the early morning hours. From in vitro exper- japonica, k = 0.154 h-', at T = 6°C (Yamashita et al. iments with T japonica, Yamashita et al. (1984) 1985). Considering all these points, it is not unreason- showed that hyperiids are able to continue feeding in able to assume that the gut evacuation rate constant darkness. Our results partly confirm this observation and passage time obtained in this study are close to the (Fig. 3) but the occurrence of all major feeding peaks real values. during daylight hours suggests that T. gaudichaudi is an active visual predator (Gibbons et al. 1992). The gut passage time obtained from the gut evacua- Daily ration tion rate experiments was equal to -7.5 h. This value, however, should be considered with great care as 2 To date, very little is known about the rate of daily independent processes may have affected the results food consumption by Themisto gaudichaudi. Accord- substantially. Firstly, gut passage time (l/k) was esti- ing to estimates of the oxygen consumed by T. gau- mated in a container filled with filtered seawater. This dichaudi adults obtained by Torres et al. (1994a), the may have caused an underestimation of llk values Minimum Food Intake (MFI) that covers the demands because Themisto gaudichaudi adults can retain food for respiration only is -1.71 % of the body dry weight. in their stomachs for period longer than 24 h in the This value is very similar to the MFI levels of 1.07 to absence of food (Gibbons et al. 1992). Secondly, we did 2.16% of body carbon calculated by Ikeda & Mitchell not estimate the rate of destruction of the secondary (1982), and slightly higher than the respiration losses pigments in the gut during digestion. Up to 50 to 90 % (1.0 to 1.2% of body weight) obtained for the amphi- of total pigments ingested by herbivorous zooplankton pod T. libellula during long-term starvation (Persy may be broken down in the gut to non-fluorescent 1993). The only thorough study of this type has been end-prod.ucts (Lopez et al. 1988, Mayzaud & Razouls undertaken with T compressa (formerly T. gaudi- 1992, Perissinotto & Pakhomov in press). In contrast to chaudi] in the northern hemisphere (Sheader & Evans the first problem, this may result in an overestimation 1975). Daily rations derived from these data range of the gut passage time. Unfortunately, the magnitude from 2.7 to 6.0% of body dry weight. In a study of the of these 2 separate factors is unknown and it is not in situfeeding ecology of T japonica (Yamashita et al. ~mpossiblethat they may compensate for each other. 1985), daily rations ranged between 4.3 and 10.9% Also, the level of gut pigment destruction in 7. gau- (mean SD = 8.64 * 1.92%)of body dry weight. dichaudi adults may even be negligible because the The in situ daily ration estimated in our study (6.3% digestive enzyme system of this carnivorous stage of of body dry weight) is at least 3.7 times higher than the the species is probably not capable of digesting plant MFI level and comparable to the values derived from pigments. A comparison of the k-value obtained in our other species of the genus Themisto. This estimate is study, 0.133 h-', with those of other crustaceans also in good agreement with the daily rations calcu- Pakhomov & Perlssinotto: Trophodynamics of Themisto gaudichaud~ 99

lated using the total energy requirements approach effect on the seasonal stock of fish eggs and larvae (7.1% of body dry weight). Rations obtained from the (Siegfried 1965, Sheader & Evans 1975, Yamashita et predation experiments were, however, generally al. 1985). Studies on the predation impact by T. gau- higher, ranging from 8.5 to 21.8% of body dry weight. dichaudi in the Benguela Upwelling System have These are more comparable to the daily rations shown rates in the range of 4 to 38.5% of copepod obtained in vitro with T. japonica, of 15.4 to 24.2% standing stock removed per day (Gibbons et al. 1992). (mean 20.3%) of dry body weight (Yamashita et al. These estimates were obtained under the assumption 1984). that densities of both prey and amphipods covary The lowest of these estimates, 8.5% of dry body closely in time and space. The maximum abundances weight, exceeds the in situ ration by 25 % only and was of T. gaudichaudi used ranged between 2.3 and derived from an experiment involving an initial prey 22.4 ind. m-3. In our study, however, maximum densi- concentration of 354 ind. I-'. According to data derived ties of T. gaudichaudi adults, averaged over the 0 to from Bongo and WP-2 net catches (H. Verheye 200 (100) m layer, never exceeded 1.5 ind. m-3 Conse- unpubl.), zooplankton abundances averaged over the quently, the predation impact was never higher than 0 to 200 or the 0 to 100 m layers ranged from 0.08 to 2.1% of the total mesozooplankton standing stock. In 0.52 ind. 1-' (mean 0.2 ind. 1-l). These figures are at terms of daily secondary production consun~ed,how- least 2 orders of magnitude lower than the concentra- ever, this impact was substantial and locally reached tion used in the experimental jar. However, the fact levels of -?0%, suggesting that T. gaudichaudiadults that daily ration levels derived from in vitro and in situ are at times able to control the mesozooplankton stock methods are reasonably similar shows that averaged in the vicinity of South Georgia. Assuming an assimila- zooplankton abundances do not reflect the microscale tion efficiency of -80% (Vinogradov & Shushkina patchiness typical of naturally occurring zooplankton 1987), the contribution of T gaudichaudito the down- communities (Navorotsky & Zadonskaya 1991, ward flux of faeces could account for <0.1 to 14 %) of Nemirovsky et al. 1991). Indeed, to survive on low the mesozooplankton daily secondary production. daily rations of 6.3 to 7.1% of body dry weight, The main results of this study show that the hyperiid Themisto gaudichaudi would need to feed on prey Themisto gaudichaudi is an important predator of concentrations at least an order of magnitude higher mesozooplankton and is able to contribute signifi- than those observed in the net samples. cantly to the vertical flux of biogenic carbon. Although, Thernistogaudichaudi is an active predator exhibit- these estimates are quite representative of the average ing daily rations that depend on the prey concentra- conditions in the natural environment, the potential tion. The species is adapted to sharp fluctuations in impact of T gaudichaudi may attain levels much food concentration and is able to starve for prolonged higher than these when the degree of covariance periods (Gibbons et al. 1992, Persy 1993). T. gaudi- between prey and hyperiid is particularly high, both in chaudi is, therefore, capable of satisfying its feeding time and space. demands very rapidly (Gibbons et al. 1992) when aggregations (even microscale swarms) of the appro- Acknowledgements. This study was funded by the South priate prey type are met. It has also been suggested African Department of Environmental Affairs and Tourism and by Rhodes University ((;rahamstown).We are very grate- that consumption becomes independent of prey con- ful to the officers and crew of the RV 'Africana' and to many centration when prey density exceeds a high threshold colleagues from the Sea Fisheries Research Institute (SFRI. (Yamashita et al. 1984, Gibbons et al. 1992). Thus, Cape Town) for their assistance at sea. We are also indebted daily rations of T gaudichaudi can theoretically be to Dr H. Vei-heye of SFRI for granting permission to use unpublished mesozooplankton abundance and biomass data. even higher than those estimated in situ in our study because of the high degree of prey/predator covariance. LITERATURE CITED

Atkinson A, Peck JM (1988) A summer-winter comparison of Predation impact zooplankton in the oceanic area around South Georgia. Polar Biol 8:463-433 Bost CA, Koubbi P, Genevois F, Ruchon L, Ridox V (1994) Since Themisto gaudichaudi adults can consume Gentoo penguin Pygoscelis papua diet as an indicator of prey at rates equivalent to at least 6.3% of their body planktonic availability In the Kerguelen Islands. Polar Biol dry weight per day, they could have a significant 14:147-153 impact on the local mesozooplankton communities, Chechun IS (1984) Feeding and ti-ophic relationships of some fish from subantarctic waters of the Indian Ocean. Proc especially in regions where hyperiids occur in great Zoo1 lnst (Leningrad) 127:38-68 (in Russian) abundances. It has been shown that some species of Conover RJ, Durvasula R, Roy S, Wang R (1986)Probable loss the genus Themisto can have a marked detrimental of chlorophyll-derived pignicnts during passage through 100 Mar Ecol Prog Ser 134: 91-100, 1996

the gut of zooplankton and some of the consequences Periss~nottoR, Pakhomov EA (in press) Gut evacuation rates Limnol Oceanogr 31~878-887 and pigment destruction in the Antarctic krill Euphausia Dam HG, Peterson WT (1988) The effect of temperature on superba. Mar Biol the gut clearance rate constant of planktonic copepods. J Permitin YE, Tarverdieva M1 (1972) The food of some Ant- Exp marBiol Ecol 123:l-14 arctic fish in the South Georgia area. J Ichthyol 12(1): Everson I, Ward P (1980) Aspects of Scotia Sea zooplankton. 104-114 B~olJ Linnean Soc 14.93-101 Persy JA (1993) Energy consumption and metabolism during Fort~erL, Le Fevre .I, Legendre L (1994) Export of blogenic starvation in the Arct~chyperiid amphipod Themisto libel- carbon to fish and to the deep ocean: the role of large lula Mandt. Polar Biol 131549-555 plankton~cmicrophages. J Plankton Res 16:809-839 Robertson AA, Alexander DGW, Miller DGM (1981) Modified Gibbons MJ, Stuart V, Verheye HM (1992) Trophic ecology of collapsible opening and closing midwater trawls (RMT-8 carnivorous zooplankton In the Benguella. S Afr J Mar Sci and RMT-2). Fish Bull S Afr 14:103-113 12.421-437 Rodhouse PG, White MG, Jones MRR (1992)Trophic relations Hardy AC, Gunther ER (1935) The plankton of the South of the cephdlopod Martialia hyadesi (Teuthoidea: Ommas- Georgia whaling grounds and adjacent waters, 1926- trephldae) dt the Antarctic Polar Front, Scotia Sea. Mar 1927. Discovery Rep 11:l-456 Biol 114:415-421 Hopkins TL (1985) Food web of an Antarctic midwater eco- Schneppenheim R, Weigmann-Haass R (1986) Morphological system. Mar Biol89:197-212 and electrophoretic studies of the genus Themisto (Am- Ikeda T (1989) Are Antarctic zooplan.kton metabolically more iphipoda: Hyperiidea) from the south and north Atlantic. cold-adapted than Arctic zoopl.ankton~An intra-generic Polar Biol 6:215-225 comparison of oxygen consumption rates. J Plankton Res Sheader M, Evans F (1975) Feeding and gut structure of 11:619-624 Parathemisto gaudichaudii (Guerin) (Amphipoda, Hyperl- Ikeda T, Mitchell AW (1982) Oxygen uptake, ammonia excre- idea). J Mar Biol Ass UK 55:641-656 tion and phosphate excretion by krill and other Antarctic Siegfried WR (1965) Observations on the amphipod Parath- zooplankton in relation to their body size and chemical emisto gaudichaudii (Guer.) off the west coast of South composition. Mar Blol 71:283-298 Africa Zoo1 Afr 1(1).339-352 Kane JE (1966)The dlstr~butionof Parathemisto gaudichaudii Simard Y,Lacroix G, Legendre L (1985) In situ twilight graz- (Guer.),with observations on its life-history in the 0' to 20°E ing rhythm during d~elvertical migrations of a scattering sector of the Southern Ocean. Discovery Rep 341163-198 layer of Calanus hnrnarchicus. Limnol Oceanogr 30: Kane JE (1967) Observations on the moulting and feeding of 598-606 a hvperiid amphipod. Crustaceana 6:129-132 Strelnikova VM (1989) Relationship between body weight Kock KI l, Wilhelms S, Everson I,Grijger J (1994) Variations in and length, and caloric content of the Antarctic hyperiid the d~etcomposition and feeding intensity of mackerel Parathemisto gaudichaudii. Hydrobiol J (Kiev) 25(1): icef~sh Champsocephalus gunnari at South Georgia 107-109 (in Russian) (Antarctic).Mar Ecol Prog Ser 108:43-57 Strickland JDH, Parsons TR (1968) A pracbcal handbook of Lava1 P (1980) Hyperiid amphipods as crustacean parasitoids seawater analysis. Bull Fish Res Bd Can 167:l-311 associated with gelatinous zooplankton. Oceanogr Mar Sugisaki H, Terazaki M, Wada E, Nemoto T (1991) Feeding BiolA Rev 18:ll-56 habits of a pelagic amphipod, Themisto japonica. Mar Biol Lopez MDG, Huntley ME, Syles PF (1988) Pigment destruc- 109:241-244 tion by Calanus pacificus: impact on the estimation of Torres JJ, Aarset AV, Donnelly J, Hopk~nsTL, Lancraft TM, water column fluxes. J Plankton Res 10:715-734 Ainley DG (1994a) Metabolism of Antarctic micronektonic Mackas D, Bohrer R (1976) Fluorescence analysis of zoo- crustacea as a function of depth of occurrence and season. plankton gut contents and an investigation of die1 feeding Mar Ecol Prog Ser 113:207-219 patterns. J Exp Mar Biol Ecol 25177-85 Torres JJ, Donnelly J, Hopkins TM, Lancraft TM, Aarset AV, Madin LP, Harbison GR (1977) The associations of Amphi- Ainley DG (199413) Proximate composition and overwin- poda Hyperiidea with gelatinous zooplankton. I. Associa- tering strategies of Antarctic micronektonic crustacea. tions with Salpidae. Deep Sea Res 24:449-463 Mar Ecol Prog Ser 113:221-232 Mayzaud P, Razouls S (1992) Degradation of gut pigment dur- Tseitlin VB (1988) Generative production of water animals. ing feeding by a subantarctic copepod: importance of Oceanology 36:493-497 feeding history and digestive acclimation. Limnol Ocea- Vinberg GG (1956) Rate of metabolism and food require- nogr 37:393-404 ments of fish. White Russia University Press, Minsk (in Navorotsky VV, Zadonskaya TA (1991) Fine-scale vertical Russian) structure and mesozooplankton in the Subarctic Front Vinogradov ME, Shushk~naEA (1987) Functioning of the (megapolygon). Rep Russ Acad Sci 316(4):983-987 (In planktonic communities in the epipelagic zone of the Russian) ocean. Nauka Press, Moscow (in Russian) Nemirovsky MS, Piontkovsky SA, Seregin SA, Korshenko AN Williams R, Robins D (1979) Caloric, ash, carbon and nitrogen (1991) Micro-scale distribution of zooplankton: adequate- content in relation to length and dry weight of Parath- diffusively processes and behaviour of individuals. In: emisto gaudicha~~dil(Amphipoda: Hyperiidea) in the Commercially important lnvestigations of plankton (col- north east Atlantic Ocean. Mar Biol 52 247-252 lected papers). VNIKO Publishers, Moscow, 1:100- 11 1 (in Yamashlta Y, Aoyama T, ffitagawa D (1984) Laboratory stud- Russian) ies of predation be the hyperiid amphipod Parathemisto Nemoto T, Yoo K1 (1970) An amphipod (Parathemisto gaudi- japonica on larvae of the japanese sand-eel Ammodytes chaudil) as a food of the Antarctic Sei Whale. Sci Rep personatus. Bull Jap Soc Scient Fish 50(7):1089-1093 Whales Res Inst Tokyo 22.153-158 Yamashita Y, Kitagawa D, Aoyama T (1985) A field study of Pakhomov EA. McQuaid CD (1996) Distribution of surface predation of the hypenid amphtpod Parathemisto japonlca zooplankton and seablrds across the Southern Ocean on larvae of the japanese sand-eel Amrnodytes person- Polar Blol 16 (in press) atus. Bull Jap Soc Sc~entFish 51(10):1599-1607

This article was submitted to the editor Manuscript first received: August 10, 1995 Revised version accepted: October 25, 1995