Are bacteria active in the cold pelagic ecosystem Barents Sea ? T . FREDE THINGSTAD and INGRID MARTINUSSEN of the Thingstad, T. F. & Martinussen, I. 1991: Are bacteria active in the cold pelagic ecosystem of the Barents Sea? Pp. 255-266 in Sakshaug, E., Hopkins. C. C. E. & 0ritsland. N. A. (eds.): Proceedings of the Pro Mare Symposium on Polar Marine Ecology, Trondheim, 12-16 May 1990. Polar Research l O ( 1 ) . Bacterial biomass and activity indicators have been studied at low water temperatures (-1.9 t o +4"C) in the Barents Sea. Strong responses by indicators of bacterial activity. such as hydrolytic enzyme and substrate uptake potentials, were observed in association with the development of phytoplankton blooms. At late successional stages of blooms, observation by cpifluorescence microscopy revealed heavy bacterial colonisation of detrital matter, in particular of senescent colonies of Phaeocystis poucherii. Based on the retention of bacteria on filters of 1 pm pore size, up to 55% of the bacterial population was estimated to be attached to organic aggregates in some cases. Based on thymidine incorporation and a conventional conversion factor, bacterial generation times as short as one day were estimated at temperatures below zero. Changes in substrate availability governed by the successional stages of the planktonic ecosystem seem t o be more important as controlling factors for bacterial growth than the low temperatures of the Barents Sea. T. Frede Thingsfad and 1. Martinussen. Deparfment of Microbiology and Plant Physiology, University of Bergen, Jahnebakken 5. N-5007 Bergen. Norway. Introduction Pelagic food webs are dynamic systems where both transient responses and equilibrium states depend upon the interactions between a large number of processes. If these processes differ in temperature sensitivity, the balance between them may shift, and, as a result, food webs in polar regions may theoretically behave differently than those of temperate regions. One such mech- anism was proposed by Pomeroy & Deibel(l986) when considering bacterial degradation of organic material. Based on data indicating low bacterial activity below +4"C, they made the suggestion that bacterial processes are of less importance in cold than in temperate waters. One of the significant aspects of such a mechanism is the potential effect of temperature on the relative importance of bacterial dissolution/degradation of particulate matter, versus metazoan feeding on these particles. With the large variation expected in time and space for both of these processes, and with the technical difficulties associated with their measurement, a direct comparison of the relative importance of these two processes for different regions is not easily obtained. If, however, bac- terial growth in cold waters is severely inhibited, phytoplankton blooms in cold waters would not be expected to be followed by any strong responses in bacterial biomass or activity. A pre- dominance of metazoan consumption prior to bacterial invasion of detrital particles would also make the observation of colonised particulate material unlikely. In the polar and meltwater regions of the Bar- ents Sea, phytoplankton blooms occur at water temperatures close to freezing point (Skjoldal & Rey 1989), and the subsequent processes of DOM and detrital POM transformation must also occur at this low temperature. These blooms are often dominated by the colony forming prymnesio- phyte Phaeocystis pouchetii (Skjoldal & Rey 1989). In temperate areas, Phaeocystis blooms have been shown to leave large amounts of organic material in the water (Eberlein et al. 1985), and after a delay of 5-10 days, the blooms are followed by a development of planktonic bac- teria (Billen & Fontigny 1987). If low-tem- perature inhibition of bacterial degradation is important, one would expect bacterial responses to Phaeocystis blooms in the Barents Sea to be feeble, more delayed or even absent. Processes related to the bacterial dissolution and degra- dation of particles involves the hydrolysis of macromolecules, the uptake of monomers, and the formation of bacterial biomass. We have found that bacterial abundance, activity poten- 256 tials for hydrolytic enzymes and uptake potentials for monomers, as well as the incorporation rate of thymidine. all increase i n response to phy- toplankton blooms occurring at temperatures between -1 and +4OC i n the Barents Sea. I n senescent blooms. massive bacterial invasion of colonies was also observed. Little support was therefore found for a bacterial response quali- tatively different from that described in temperate waters. T . F. Thingstad & I . Martinussen Materials and methods Sampling Samples were collected by a rosette sampler with Niskin water bottles attached to the CTD sonde, or, when large volumes were required, by 30L Niskin bottles shortly after the CTD cast. Geo- graphical position of stations are given in Fig. 1. The stations are from different cruises: Stations 14. 24. 31. 52 are from cruises with coast guard vessels K / V ANDENES and K ~ V S E N J A in April 1986. 80 z 0 78 76 w 0 2 - 74 l- J a Stations 733. 890 and 961, with R/V G . 0. SARS in June 1984. August 1985 and June 1987, respect- ively. Temperature dependence of proteases was measured on a sample of brown ice collected in April 1986 and thawed at an ambient temperature of 10°C. Incubations where performed at close to in situ temperature in a thermostatted water cooled incubator, or, for temperature depen- dence studies, in a temperature gradient incu- bator described by Eriksen & G o k s ~ y r (1976). Microscopic counts Heterotrophic and autotrophic pico- and nano- plankton were counted by epifluorescence micro- scopy according to Martinussen & Thingstad (1991). Due t o a double staining with primuline and DAPI and t h e alternation between blue and UV-exitation in the epi-fluorescence microscope, this procedure allows identification of autotrophs by chlorophyll autofluorescence, positive identi- fication of cells based on the existence of a DAPI- stained nucleus. and the visualisation of the whole J 890 24, ,961 v i e 14 31 0 Bjerneya I 7 2 I Norway 70 - I I I I 0 20 30 40 I L O N G I T U D E O E Fig. I . Positions of stations in the Barcnts sea. 3 Are bacteria active in cold waters? 257 cell by the primuline stain. The primuline stain also allowed the visualisation of the mucus of Phaeocystis colonies. Samples were fixed with 2.5% final concentration of borate-buffered for- malin. As a standard procedure, a 5 m l sample was filtered on a 0.2 pm pore-sized polycarbonate filter for bacterial counts and between 15 and 50 ml, dependent upon expected population den- sity, on a 1 pm filter for counting of eucaryotes and retained bacteria. Bacterial cell volumes were calculated from linear dimensions determined in the microscope at lOOOx magnification using a calibrated graticule (New Porton G12, Graticules Ltd., England) in one ocular. Enzymatic activity potentials Activity potentials for protease and Fglucosidase were determined according to Somville & Billen (1983) and Somville (1984) using a Perkin-Elmer LS5 fluorescence spectrophotometer. Thymidine incorporation Thymidine incorporation into cold TCA-pre- cipitable material was determined according to Fuhrman & Azam (1982) modified by increasing the concentration of added thymidine to 10 nmole I-'. Conversions from thymidine incor- porated to estimates of cell production rates are based on a conversion factor of 2.4 10IR cells mol-' as suggested by Fuhrman & Azam (1982) for offshore waters. Generation times are computed as In2 x cell density (cell production rate)-'. Uptake potentials of I4C-labelled glucose and amino a c i h Uptake potentials for monomers were measured by adding 37 kBq of 14C-labelled glucose (Amer- sham CFB.96) or a I4C-labelled amino acid mix- ture (Amersham CFB.104) to 10ml samples of seawater. Final concentrations of glucose and amino acids were 0.37 and 0.38 pmol I - ' , respect- ively. The different amino acids in the mixture had different specific activities, and conversion from DPM to nmoles are based on mean specific activity of the mixture (68 E q - mmol-I). After 5 h incubation, the samples were filtered on 0.2 pm cellulose-nitrate membrane filters. The filters were washed twice with filtered seawater. dried, and counted by scintillation counting. Chlorophyll, NOj, in-situ fluorescence and light These data were obtained from the sampling pro- gramme of the Marine Research Institute, Bergen. Chlorophyll a concentrations were measured fluorometrically after acetone extrac- tion of filtered samples, nitrate was measured by autoanalyser as described in Foyn et al. (1981), fluorescence profiles were obtained using a sub- mersible fluorometer (Q-fluorometer). An excep- tion to this is Station 961, where profiles of in situ fluorescence, light, and a, profiles are obtained from B. G . Mitchell, Scripps Institution of Ocean- ography, and were collected using a bio-optical rig (Biospherical Instruments) equipped with sensors MER-12, MER-08, and MER-53. Results Ice-edge bloom, April I986 The bacterial development during the first phase of the ice-edge bloom was followed on a cruise Fig. 2. Station 31. April 1986. Profilcs of salinity (broken line). tcmpcrature (solid linc). and protcasc activity potential (solid circles). Development of bacterial activitics occurring in thc cold meltwatcr layer on top of warmer Atlantic Water masses. 258 T. F. Thingstad & I . Martinussen ~ A I I 1 I I B 6 11 16 21 e 11 16 2 1 1 4 24 31 Date April 1986 l 4 24 31 52 52 W t i o n n r . Fig. 3. Scqucncc of stations during icc-cdgc bloom. Samplcs from 10 m dcpth A . Chlorophyll (I (opcn triangles). NO, (open squares). and tcmpcraturc (opcn circlcs) B. Bacterial abundancc (solid circlcs). bactcrial biovolumc (solid squarcs). and generation times estimated from thymidinc incorporation (solid triangles). C. Protcase activity potential (solid circles). uptake potential f o r amino acids (opcn circles). D. /?-glucosidasc aetlvity potcntial (solid trianglcs). and glucose uptakc potcntial (open trianglcb). Are bacteria active in cold waters? 259 with coast guard vessel K/V ANDENES in April 1986 in the area west of Sentralbanken in the Barents Sea (Fig. 1 ) . The meltwater layer in the area extended down to about 50m (Fig. 2). Return of the ship to the same area with intervals of a few days allowed the ordering of samples in a time sequence spanning 13 days, permitting an indication of how the bloom in the area devel- oped, although bo strict time sequence of the development in a well-defined water body was possible. During the period of observation, the increase in chlorophyll levelled off, and nitrate was consumed. A slight heating of the meltwater Layer occurred, with water temperatures at 10 m reaching -1.O"C at the end of the period (Fig. 3A). Bacterial abundance increased by a factor of 3 in numbers and 5.6 in total biovolume (Fig. 3B), the main increase in bacterial bio- volume occurred at the end of the period. All measurements of bacterial activity indicated a transition to a phase of active bacterial growth at the end of the period. The enzymatic activity potentials increased throughout the period with factors of 60 and 40 for 6-glucosidase and protease respectively. Due t o a temporary failure of the scintillation counter, uptake measurements are only available for Stations 31 and 52, but glucose and amino acid uptake potentials both showed the same rapid increase as the enzyme activity potentials (Fig. 3C, D). The bacterial response was restricted to the cold water layer with no response in the warmer Atlantic Water below as illustrated by the profile of protease activity potential in Fig. 2. Estimates of bacterial gen- eration times from thymidine incorporation indi- cated a decrease in mean generation time of the bacterial population from 13 to 4 days between the 8th and the 13th day of the period (Fig. 3B). Subsurface chlorophyll maximum Station 961, in the central Barents Sea, June 1987, is an example of a later successional stage where a sharp subsurface chlorophyll a maximum has developed at the pycnocline (Fig. 4A) between the meltwater layer and the underlying Atlantic Water. The autotrophic community was com- pletely dominated by Phaeocystis. The profile of bacterial abundance had a sharp peak very similar to the distribution of Phaeocystis cells (Fig. 4B), but with a maximum of 2 . lo6 ml-' shifted slightly downwards in the water column relative to the peak in abundance of Phaeocystis. Activity poten- tials of hydrolytic enzymes and uptake of glucose and amino acids all indicated the same type of depth profiles with peaks associated with the upper part of peak in bacterial abundance (Fig. 4C, D). At this station, colonies of Phaeocystis were still relatively free of bacteria, and maximum retention of bacteria on 1 pm filters was 10% at 10m. At the fluorescence maximum 3% were retained. Bloom in water without well-defined pyknocline Station 733 in the eastern Barents Sea, June 1984, represents a situation with temperatures below zero at all depths. There was no distinct pycno- cline, but a gradual increase in ut downwards. (Fig. 5A). The in situ fluorescence increased down to 20 m which corresponded to the bottom of the photic zone. A broad maximum in the fluor- escence profile extended down to about 60 m with a long "tail" extending to the lowest measured depth at 90 m. Chlorophyll at the depths of maxi- mum fluorescence corresponded to 3.8 pg Chl I - ' . Estimates of bacterial generation time had two minima, 1.2 and 1.5 days, at the surface and at 30m respectively (Fig. 5B). Heavily invaded Phaeocystis colonies were observed and bacterial retention on 1 pm reached a maximum of 40% at 30m depth (Fig. 5B). Post-bloom situation. August 1985 Station 890 in the central Barents Sea, August 1985, represents a situation with low values of in situ fluorescence throughout the water column, combined with nutrient depletion above the pycnocline (Fig. 6A). Except for a maximum of 0.9 pg Chl * I - ' at 20 m, values were below 0.4 pg Chl - I - ] . Water temperature above the pycno- cline at 10m were +4"C, decreasing to -1.8 in the underlying polar waters. The phytoplankton community was dominated by a small, unidenti- fied picoplanktonic species approximately 1.5 pm in diameter and by Dinobryon spp. The fastest bacterial growth was found at the bottom of the pycnocline (20m) (Fig. 6B) where a bacterial generation time of 1.0d was estimated. At all depths of this station, aggregates with bacterial colonisation were observed, a feature particularly prominent at 20m. This was reflected in a high fraction (55%) of the bacterial population being retained on 1 pm polycarbonate filters. The extreme maximum of bacterial retention in the 260 T. F. Thingstad & I . Martinussen T e m p e r a t u r e " C - 1 . 0 - 0 5 0 0.5 1 0 Phaeocvstis cells 0 1 2 i u m i n - situ :Iuorescence re1 units 0 1 2 3 4 1 B a c t er I Q 0 2 106in1-1 2 7 6 21 20 0 Gt Amino acid upt.n h e 0 G l u c o s e u p t a k e A nm0l.l" .h' 0 0 1 0 2 o 3 nrnol.l".h' o 002 0 0 4 006 0 O R I I I P r o t e a s e a c t i v i t y p o t e n t i a l nmol q' 41.' - - > I C 0 Glucosidase act. p o t e n t i a L A 0 0 . 5 1 . o 1 . 5 nrnot .I-' .ti1 D Are bacteria active in cold waters? 261 Fluoresc. rehnits Nitrate amol-l-' 0 2 k 6 8 1 0 I"" ~ , 1 1 1 - 2 6 21.6 21.8 1 0 Temp o(: oi 0 25 E I w 0 50 15 100 Bad. abundance 1O'mI-l . Gemtime d 8 I I I 20 40 60 % retained Fig. 5. Bloom in water without well-defined pycnocline. Station 733, June 1984. A . Profiles of fluorescence (solid line). tem- perature (dotted line). u, (broken line) and nitrate (open squares). B . Bacterial abundance (solid triangles). generation time estimates (solid squares), and percentage of bacterial population retained on 1 pm polycarbonate filters (solid circles). vicinity of the pycnocline was specific for this station. The origin of the organic aggregates at this station is unknown. At other stations, however, senescent colonies of Phaeocystis p o u - chetii heavily colonised by bacteria were easily recognisable. Classification of the water samples according to whether the phytoplankton com- munity was dominated by Phaeocystis, the unidentified picoplankton sp., a mixture of these, or dominated by other species demonstrated a clear correlation between the Phaeocystis/ picoplankton communities, and a high retention of bacteria by 1 pm filters (Fig.7). Temperature sensitivity of protease The Arrhenius plot for hydrolysis of the artificial substrate by the natural mixture of proteases in brown sea ice was linear (R2 = 0.996) in the inter- val - 1.5 (lowest temperature investigated) to 21°C (Fig. 8). The slope of the linear part cor- responds to an activation energy of 12.1 kcal . mole-' which corresponds to a Qlo in the range -1.5 to 8.5"C of about 2.2. Discussion Bacterial abundances and activities, not drast- ically different from those of temperate waters, may develop in the Barents Sea, even at subzero temperatures. This is a conclusion in accordance with results previously reported by investigators in the Antarctic Ocean (Hodson et al. 1981; Han- son & Lowery 1985). We found, however, a large range in values of thymidine uptake, from less than 0.004 nmole . I - ' . day-' in samples taken in late winter before the onset of the ice edge bloom to 0.26 nmole I-' . day-' in connection with deep chlorophyll maxima in August. Our sampling did not allow a strict deter- mination of the timing between the peak of the ice-edge phytoplankton bloom and the bacterial response. The sequence of stations in April 1986 (Fig. 3) did, however, suggest that the response was initiated even before the exhaustion of the nitrate. No indications were found of a delay substantially longer than the 5-10 days reported for temperate waters by Billen & Fontigny (1987). At later stages of the bloom, a close correlation between the profile of bacterial abundance and the deep fluorescence maximum could be observed (Fig. 4), with bacterial abundances up Fig. 4. Deep fluorescence maximum, station 961. June 1987. A . Fluorescence (solid line), temperature (dotted line) and u, (broken line). B . Cell counts of Phoeocysfis (solid triangles), and bacteria (solid squares). C. Protease activity (solid squares) and amino acid uptake (open squares) potentials. D . b-glucosidase activity (solid triangles), and glucose uptake (open triangles) potentials. 262 T. F. Thingstad & I . Martinussen F I u o r escen ce .. r e I. un i t s iri 5 10 0 B a, 26.5 2 2 0 27.5 28.0 J I I I N i t r a t e pM 0 5 10 1 5 10 - 20 - 30- E I- = : LO- 5 0 - 60 7 - 5 0 5 T e m p , O C I I 1 1 0 20 40 60 ‘10 R e t a i n e d 0 1 50 - 40. 30 20 10 ( 4 33 11 5 0 Ph Pi Ph+Pi PhytopLcommuni t y fig. 7. Retention of bacteria on 1 pm polycarbonate filters. Water samples from cruise in August 1985 classified according to dominance of phytoplankton community by Phaeocysrk (Ph). picoplankton (Pi). a combination of these (Ph + Pi). or by other groups (0). Bars and vertical lines indicate mean and range. rcspectivclv. Number of water samples in each category givcn abobc each bar. F q 6 Post-bloom situation. Station XYO. August IY8.5 A Profiles of fluorescence (solid line), temperature (dotted line). u, (hroken line), and nitrate (open squares). B. Bacterial ahun- dancc (solid triangles). generation time cbtimatcs (solid squarca), and retention on I u m polycarbonatc filters (solid circles). Are bacteria active in cold waters? 263 3.5- 3.0- 2.0 - 2 ’ o - Fig. 8. Arrhenius plot of the rate of hydrolysis of L-leucyl-pnaphtyl- amide by proteases from a sample of brown ice . \ 8 \ Temp I OC 30 I 21 I I 1 3 I 5 ‘I \. to 2 + lo6 ml-l, similar to levels observed in coastal temperate waters (Azam et al. 1983). Our gen- eration time estimates are based on a conversion factor between thymidine incorporation and cell production rate which is not determined for the actual environment(s). Since the factor is known to vary (Fuhrman & Azam 1982; Riemann et al. 1987), our generation time estimates must be taken as indicative only. Some of the higher values obtained for the conversion factor seem, however, t o be associated with the addition of thymidine at concentrations below 10 nmole * I-’, while lower values have been determined in coastal waters (Riemann & Bell 1990). The fastest growth with generation times about 1 day are, however, comparable to generation times esti- mated in temperate waters. As an example, Rie- mann et al. (1984) report values based on thymidine uptake between 17 and 91 hours through a die1 cycle in a Danish coastal environ- ment. In laboratory studies, Harder & Veldkamp (1971) reported a maximum growth rate of 0.073 h-l at -2°C for one of the species inves- tigated (Pseudomonas L12), corresponding to a generation time of 9.5 hours. This is below the generation times estimated by us, and dem- onstrates the feasibility at low temperatures of generation times that are short when viewed in an ecological context. The maximum protease activity potential observed at Station 52 (Fig. 3C) was 1 . 1 nmole * I-’ . min-l. Using the tempera- ture dependence determined, this corresponds to an activity of about 3 nmole I - ] . min-l around 15°C. This is comparable to typical values reported by Fontigny et al. (1987) for coastal seawater in the North Sea, although maximum values found in this more eutrophic environment exceeded 20 nmole . I-’ . min-l. From the visual observation of how bacterial microcolonies had formed on senescent Phaeo- cystis colonies, we felt fairly convinced that the colonisation was accompanied by active growth of attached bacteria. We have, however, no esti- mate of the relative proportion between those bacteria that were primary invaders and their offspring on the mucous material. In situations with high mucus content of the water, the reten- tion of bacteria by 1 pm filters may overestimate the proportion of attached bacteria due to entrap- ment of bacteria in the mucus during the filtration process. The observation of colonies without attached bacteria during early stages of the blooms do, however, show that fresh mucus did not collect bacteria during the filtration process. While we feel confident that our observation shows that detrital particulate matter is colonised, we can not determine to what degree the material had been respired. A possible difference in the temperature responses of growth and respiration has been suggested (Christian & Wiebe 1974). A large temperature dependence for respiration could imply a large change in consumption rate 264 of organic matter, without this being revealed in measurements related to changes in bacterial biomass. Temperature has been shown to influence most of the important growth parameters for bacteria. In a study of Aerobacter aerogenes (25 to 40°C). Topiwala (1971) found that parameters such as the half saturation constant for growth and endogenous metabolism were temperature dependent. Harder & Veldkamp (1967) found that a decrease in temperature was compensated for by an increase in the concentration of RNA and respiratory enzymes. Bacterial degradation of organic matter within an ecosystem context is, however, not simply a function of the physio- logical properties of the bacteria; degradation may also be strongly controlled by trophic inter- actions such as competition, predation, and remineralisation (Pengerud et al. 1987), and therefore by the temperature dependence of such processes. In a system where bacterial growth rate is controlled by organic or inorganic substrate limitation. moderate changes in maximum growth rate with temperature may be unimportant or compensated for by the dynamics of the system. An example is perhaps a possible compensation for low temperature by high substrate con- centrations as suggested by Pomeroy & Wiebe (1988). Supporting evidence for this was found in investigations of amino acid concentrations and bacterial activity in the arctic (Pomeroy et al. 1990). If extracellular hydrolysis is the rate limiting step of macromolecular degradation as suggested by Billen (1988), the temperature dependence of this process is particularly important. The exten- sion of the linear part of the Arrhenius plot for hydrolysis by Barents Sea proteases t o tem- peratures below 0°C (Fig. 8) demonstrates that enzymes functioning normally at low tempera- tures have developed in this ecosystem. The exist- ence of bacteria with a complete enzymatic machinery functioning normally at low tem- peratures has been demonstrated two decades ago by Harder 81 Veldkamp (1971) who studied a group of Pseudomonas spp. for which the linear part of the Arrhenius plot for maximum growth rate extended down t o between - 4 and -5°C. Qlo values reported for enzyme catalysed reactions in natural systems seem to be fairly uniform. The value of about 2.2, corresponding t o an activation energy of 12.1 kcal . mole-', found for hydrolysis by Barents Sea proteases, can be compared to the T . F. Thingstad & I . Martinussen value 12.0 kcal . mole-' reported for hydrolysis of casein with trypsin (White et al. 1968), to the value of 11 kcal . mole-' reported for maximum growth rate of both obligate and facultative psych- rophilic bacteria (Harder & Veldkamp (1971), or to a range of Q l o values between 1.8 and 2.3 for photosynthetic capacity of natural phytoplankton as reviewed by Harris (1980). Any extreme sensitivity of bacterial processes to low temperatures of the type suggested by Pomeroy & Deibel (1986) or Pomeroy & Wiebe (1988) does therefore not seem to be of universal validity or to be a necessary consequence of any fundamental limitations in bacterial physiology. The reason for this apparent discrepancy in results is not immediately obvious. One may speculate on the possibility that there are differences in the composition of the bacterial community in permanently and in seasonally cold waters. Polar Water masses i n the Barents Sea are characterised by temperatures below 0°C. The Atlantic inflow has a mean of 6.2"C, but this decreases rapidly east and northwards into the Barents Sea (Loeng 1989), and the Barents Sea therefore classifies as a permanently cold environment. In a one-year study in the Baltic when surface temperatures varied from 0 to 14T, Kuosa & Kivi (1989) found a peak in bacterial production following the spring phytoplankton bloom in a period during which water temperatures increased from 2 to 6°C. This was a result seemingly in qualitative accordance with our results from the Barents Sea. In experi- ments with water collected in April from thz same area, however, bacterial growth was not found when filtered samples were incubated at in situ temperature (+ 1"C), although higher incubation temperatures induced rapid bacterial growth (Autio 1990), suggesting that cold-adapted bac- teria were not present. A unified picture of how the bacterial community adapt to temperature in permanently and in seasonally cold environments seems therefore still to be lacking. Ecosystem behavior is a result not only of the interactions between biological processes but also of the interactions between biological and physi- cal processes. Perhaps the latter is a field where differences in temperature dependence of inter- acting rates is more likely t o be expected. As an example, viscosity of 35% seawater decreases by a factor of only 1.36 between 0 and 10°C (Knauss 1978). One may speculate whether this response is sufficiently small to shift the balance between sedimentation and degradation rate of particles Are bacteria active in cold waters? 265 and, as a result, influence the degree of nutrient impoverishment of the meltwater layer or the amount of organic material reaching the sedi- ments. Such a shift would constitute a third element to be considered in the discussion of whether a match in time between the spring phy- toplankton bloom and the development of mezo- zooplankton is important for the partitioning of the primary production between sedimentation and predation by metazoans in the Barents Sea ecosystem (Rey et al. 1987; Wassmann 1989). It would also effect the carbon budget of the upper ocean. Acknowledgements. - This work is part of the Norwegian Research Program for Marine Arctic Ecology (Pro Mare) and was financed by the Norwegian Research Council for Science and the Humanities and the Norwegian Ministry of the Environ- ment. We wish to thank colleagues within this programme for cooperation on the cruises and for supplying CTD, fluorescence and mineral nutrient analyses used in this article and 1. Dundas for comments on the manuscript. References Autio, R. M. 1990: Bacterioplankton in filtered brackish water cultures: some physical and chemical parameters affecting community properties. Arch. Hydrobiol. 117. 437451, Azam. F.. Fenchel, T., Field, J . G . , Gray, J . S., Meyer-Reil. L. A . & Thingstad, T. F. 1983: The ecological role of water- column microbes in the sea. Mar. Ecol. Prog. Ser. 10. 257- 263. Billen, G . 1988: Delayed development of bacterioplankton with respect to phytoplankton: A clue to understanding their trophic relationships. 4th Inr. Workshop on Measurement of Microbial Activities in the Carbon Cycle. Ceske Budejovice, Czechoslovakia. Billen, G. & Fontigny, A. 1987: Dynamics of Phaeocystis dom- inated spring bloom in Belgian coastal waters. 11. Bac- terioplankton dynamics. Mar. Ecol. Prog. Ser. 37, 249-257. Christian, R . R., & Wiebe. W. I . 1974: The effects of tem- perature upon the reproduction and respiration of a marine obligate psychrophile. Can. 1. Microbiol. 20, 1341-1345. Eberlein, K . , Leal, M. T.. Hammer. K. D. & Hickel. W. 1985: Dissolved organic substances during a Phaeocystis pouchefii bloom in the German Bight (North Sea): Mar. Biol. 89,311- 316. Eriksen. J . & Gokseyr. J . 1976: The effect of temperature on growth and cellulase (p-1 .4-endoglucanase) production by the compost fungus Chaetomium fhermophile var. dissitum. Arch. Microbiol. 110. 233-238. Feyn, L . . Hagebe, M. & Seglem. K. 1981: Automatisk analyse av nmingssalter med "on-line" databehandling. Fisken og H a w . Serie B . Fontigny, A.. Billen. G . & Vives Rego, J . 1987: Some kinetic characteristics of exoproteolytic activity in coastal seawater. Est. Coast. Mar. Sci. 25, 127-134. Fuhrman, J. A . & Azam, F. 1982: Thymidine incorporation as a measure of heterotrophic bacterioplankton production in marine surface waters: Evaluation and field results. Mar. Biol. 66, 1 W 1 2 0 . Hanson, R. B. & Lowery, H . K. 1985: Spatial distribution, structure, biomass. and physiology of microbial assemblages across the southern ocean frontal zones during the late austral winter. Appl. Enuiron. Microbiol. 49. 1029-1039. Harder, W. & Veldkamp. H. 1967: A continuous culture study of an obligately psychrophilic Pseudomonas species. Arch. Microbiol. 59, 123-130. Harder, W. & Veldkamp, H. 1971: Competition of marine psychrophilic bacteria at l o w temperatures. Antonie iian Leeuwenhoek 37, 51-63. Harris, G . P. 1980: The measurement of photosynthesis in natural populations of phytoplankton. Pp. 129-187 in Morris. 1. (ed.): The physiological ecology of phyroplankton. Blackwell. Oxford. Hodson, R. E., Azam. F., Carlucci. A. F., Fuhrman, J . , Karl. D . M. & Holm-Hansen. 0. 1981: Microbial uptake of dis- solved organic matter in McMurdo Sound, Antarctica. Mar. Biol. 6 1 , 89-94. Knauss, J. A . 19 7 8: Introduction 10 physical oceanography. Prentice Hall p. 324. Kuosa, H. & Kivi, K. 1989: Bacteria and heterotrophic flag- ellates in the pelagic carbon cycle of the northern Baltic Sea. Mar. Ecol. Prog. Ser. 53, 95-100. Loeng, H. 1989: Ecological features of the Barents Sea. In: Rey, L. & Alexander, V. (Eds.): Proc. 6th Conf. Com. Arctique Internat., 13-15 May 1985. E . J . Brill. Leiden 1989. Martinussen, I . & Thingstad. T. F. 1991: A simple double- staining method for enumeration of autotrophic and het- erotrophic nano- and picoplankton. Mar. Mirrob. Food Webs 5 , 5 1 1 . Pengerud, B.. Skjoldal, E. F. & Thingstad. T . F. 1987: The reciprocal interaction between degradation of glucose and ecosystem structure. Studies in mixed chemostat cultures of marine bacteria. algae. and bacterivorous nanotlagellates. Mar. Ecol. Prog. Ser. 35. 111-117. Pomeroy, L. R. & Deibel. D. 1986: Temperature regulation of bacterial activity during the spring bloom in Newfoundland coastal waters. Science 233, 359-361. Pomeroy, L. R.. Macko, S . A , , Ostrom, P. H. & Dunphy. J . 1990: The microbial food web in Arctic seawater: con- centration of dissolved free amino acids and bacterial abun- dance and activity in the Arctic Ocean and in Resolute Passage. Mar. Ecol. Prog. Ser. 61, 31-40. Pomeroy. L. R. & Wiebe, W. J . 1988: Energetics of microbial food webs. Hydrobiologia 159, 7-18. Rey, F.. Skjoldal. H. R. & Slagstad. D . 1987: Primary pro- duction in the Barents Sea. Pp. 29-46 in Loeng. 13. (ed.): The effect of oceanographic conditions on distribution and population dynamics of commercialfish stocks in the Barents Sea. Proc. 3rd Sovict-Norwegian Symp. Murmansk. 2 6 2 8 May 1986. Riemann, B. & Bell, R . T.. 1990: Advances in estimating bacterial biomass and growth in aquatic systems. Arch. Hydrobiol. 4 . 385-402. Riemann. B., Bjdrnsen. P. K . . Newell. S. Y . & Fallon. R. D. 1987: Calculation of bactereplankton production from measurements of 'H-thymidine incorporation. Limnol. Oce- anogr. 32, 471-476. Riemann, B.. Niclsen. P., Jeppesen. M.. Marcussen, B. & Fuhrman. J . A . 1984: Die1 changes in hacterial biomass and growth rates in coastal environments. determined by means of thymidine incorporation into DNA, frequency of dividing 266 T. F. Thingstad & 1. Martinussen cells(FDC). and microautoradiography. M a r . Ecol. f r o g . Ser. 17, 227-235. Skjoldal. H. R & Rey. F. 1989: Pelagic production and varia- bility In the Barents Sea ecosystem. Pp. 242-286 in Sherman. K . & Alexander. L (eds.): Biomass yields ondgeography of large marine ecosysrem. AAAS Selected Symposium 1 1 1 . Westview Press, Boulder. Colorado. USA. Somville. M. 1984: Measurement and study of substrate speci- ficity of exoglucosidase activity in eutrophic water. Appl. Enoiron. Microbiol. 48, 1181-1185. Somville. M. & Billen. G . 1983: A method for determining exoproteolytic activity in natural waters. Limnol. Oceunogr. 28. 1%193. Topiwala. H . 1971: Temperature relationships in continuous culture. Biofechnol. Bioeng. 13. 795-813. Wassmann. P. 1989: Sedimentation of organic matter and sili- cate from the euphotic zone of the Barents Sea. Rapp. P:u. R e u n . Corn. lnr. Explor. Mer 188. lO&l14. White A . Handler. P. & Smith, E. L. (eds.) 1968: Principles of biochemrtfrv. McGraw-Hill. New York. 233 pp.