1. INTRODUCTION Veliko and Malo Jezero (the Mljet Lakes) are natur- al phenomena on Mljet Island (Adriatic Sea). Due to its scenic beauty, ecological peculiarities, and environmen- tal values the western part of the island was proclaimed a National Park in 1960 (Fig. 1). Veliko and Malo Jeze- ro (Large and Small Lake = Mljet Lakes) are semi- enclosed depressions connected with the open sea by a narrow, shallow channel. The lakes are typical karst depressions (d o l i n e s or sinkholes), which were formed under subaerial exposure and are now submerged due to Holocene sea-level rise. Being connected with the sea, they have saline marine water and therefore are not true lakes, but can be termed marine lakes. Morphomet- ric characteristics of the Mljet Lakes were reported in VULETI∆ (1953), along with their geological setting and sediment characteristics. Benthic Foraminiferal Assemblages in a Restricted Environment - An Example from the Mljet Lakes (Adriatic Sea, Croatia) Vlatka VANI»EK , Mladen JURA»I∆ , Zlatan BAJRAKTAREVI∆ and Vlasta ∆OSOVI∆ Weak tidal currents and feeble wind influence cause expressed stratification of the water column, giving rise to temporary stagnant conditions in the marine lakes, especially in Malo Jezero (BULJAN & ©PAN, 1976; BENOVI∆ et al., 2000). Therefore the Mljet Lakes are restricted environments not only in a spatial sense, but also in the sense of stress-producing factors (SEIBOLD & BERGER, 1996). The aim of this preliminary research was to compare benthic foraminiferal assemblages, including foramini- feral morphological variations, with hydrological, sedi- mentological, and ecological characteristics of such a peculiar, restricted, unstable environment in the karsti- fied marine depressions. In spite of the worldwide dis- tribution of karstic phenomena, there are very few papers dealing with the various organisms inhabiting such areas (CIMERMAN et al., 1988; SOKA» & BAJ- RAKTAREVI∆, 1995). CIMERMAN et al. (1988) found two different foraminiferal assemblages in Veliko Jezero. In coarse- grained sediment at 30 m depth miliolids occurred in large number, while in the muddy bottom (at 42 m depth) agglutinated foraminifera were dominant, imply- ing low temperature, low oxygen content and reduced light (i b i d, p. 745). In contrast, SOKA» & BAJRAK- TAREVI∆ (1995, p. 130) have described in the same marine lake, but at 8 m depth in the >250 µm size frac- tion a typical shallow-marine foraminiferal assemblage with a prevalence of porcellaneous and agglutinated species. In a sample from 30 m depth they found that hyaline foraminifera dominate the assemblage. Salinity data in this paper are reported simply as numbers, because the salinity has been defined since the early 1980’s as the ratio between conductivity of seawa- ter and of standard KCl solution (OPEN UNIVERSITY COURSE TEAM, 1989). 2. STUDY AREA 2.1. HYDROGRAPHY Water exchange between the open sea and Veliko Jezero and Malo Jezero is driven by tidal currents. However, the average tidal amplitude in this area is only 11±0.5 cm (BULJAN & ZORE-ARMANDA, 1976). The Veliki Most (Large Bridge) strait between G EOL. CROAT. 53/2 269 - 279 3 Figs. 6 Tabs. ZAGREB 2000 Key words: Benthic foraminifera, Anoxia, Restricted environment, E l p h i d i u m, H a y n e s i n a , Mljet Island, Adriatic Sea, Croatia. Department of Geology, Faculty of Science, University of Zagreb, Zvonimirova 8, P.O.Box 332, HR-10002 Zagreb, Croatia. Abstract Benthic foraminiferal assemblages from a peculiar restricted marine environment, the Mljet Lakes (Mljet Island, Adriatic Sea, Croatia) have been studied. These lakes are drowned karst dolines, which are connected with the Adriatic Sea through a narrow, shallow channel. Occasional stagnant conditions in the marine lakes cause hypoxic and anoxic conditions in the bottom waters. Such stressed conditions are reflected in oligospecific benthic foraminiferal assem- blages with a Shannon-Wiener species diversity index (H) ranging from 0.8 to 1.0 and equitability index (E) ranging from 0.18 to 0.26, identified in samples from each marine lake. In the more dysoxic Malo Jezero, Haynesina depressula dominates an assemblage of 12 benthic foraminiferal species. In the less (and less frequently) hypox- ic Veliko Jezero, we found an Asterigerinata mamilla a s s e m b l a g e with 18 foraminiferal species. A more diverse assemblage containing 55 different benthic foraminiferal species occupies an adjacent open- sea station. Long-term salinity measurements indicate that H. depressula tol- erates higher salinity than formerly presumed (up to 38‰), and is well adapted to stressed hypoxic conditions. 270 Geologia Croatica 53/2 the Soline Channel and Veliko Jezero is only 10 m wide and 2.5 m deep. Prior to artificial enlargement in 1960 it was only 4.5 m wide and 0.6 m deep (STRA- ÆI»I∆, 1979). The Mali Most (Small Bridge) strait between Veliko and Malo Jezero is even smaller: 2.5 m wide and 0.2 m deep. Therefore the masses of water exchanged are small. Mean annual precipitation (1981- 1990) at the nearby meteorological station Goveari is 765 mm, and potential evapotranspiration is 692.6 mm/year, indicating the slightly humid climatic charac- teristics of the region. However, precipitation is con- centrated in winter, whereas during summer evapora- tion prevails (VU»ETI∆ & VU»ETI∆, 1995). There- fore, during the warm part of the year, due to excessive heating of the surface layer, large differences in density between surface and bottom waters occur. Dense bot- tom water formed in winter cannot be exchanged dur- ing summer. The concentration of organic matter in- creases in summer, and this causes enhanced bacterial degradation and oxygen undersaturation in the deepest parts of the Mljet Lakes (CARI∆ & JASPRICA, 1995). The summarized hydrographic data (Table 1, after BULJAN & ©PAN, 1976; BENOVI∆ et al., 2000) indi- cate high salinity (36-38 ‰), relatively cold water (9 - 18°C), and episodic anoxic conditions in bottom water in both depressions (in Malo Jezero and in Veliko Jeze- ro). For example, BULJAN & ©PAN (1976) report that in 1951 to 1953 there was a permanent anoxia in the bottom waters in the Malo Jezero. BENOVI∆ et al. (2000) found that dissolved oxygen saturation near the bottom in October 1997 was 4.3% in Malo Jezero and 17% in Veliko Jezero. Moreover, they registered an anoxic event in the Veliko Jezero between August 26 - 28, 1996, when below the thermocline layer (at 18.5 m) oxygen concentration rapidly decreased to 0% in the layer from 39 m to the bottom. Open marine hydro- graphic conditions were found at the Kriæ station (Figs. 1, 2). Due to vertical differences in temperature and salin- ity, the formation and depth of the seasonal thermocline differs at the locations investigated. In the open waters of Mljet Island (Kriæ station), the seasonal thermocline depth ranges between 11 and 42 m and its average depth is at 25 m. In Veliko Jezero, the thermocline is shallower (12 -22 m) and shows little vertical change. In Malo Jezero, the thermocline is the shallowest (5-12 Fig. 1 Map of Mljet Island, Croatia, showing the study area and sampling stations. m), but also the most stable and the strongest, because of large temperature differences (Fig. 2). Occasionally, the thermocline formed during summer in Malo Jezero remains stable even during winter, thus preventing bot- tom water mixing for a period longer than a year (BU- LJAN & ©PAN, 1976). 2.2. SEDIMENTS Previous sedimentological research in the Mljet Lakes (VULETI∆, 1953; SEIBOLD, 1958; CIMER- MAN et al., 1988; JURA»I∆ et al., 1995) indicated similar sediment distribution in both lakes. The deepest part of Malo Jezero is covered with muddy sediment, whereas the bottoms above 17 m contain mainly zooge- nous sand. A change in sediment type with depth can also be observed in Veliko Jezero, but it is not as regu- lar as in Malo Jezero. The shallower parts of Veliko Jezero are covered with biogenous carbonate sands, whereas in the deeper parts (42 m) clay minerals and authigenic pyrite are present. In sediment cores taken from both Malo Jezero and Veliko Jezero, microlamination is common. The lami- nae have been interpreted either as annual layers (var- ves) as a consequence of seasonal anoxic conditions (SEIBOLD, 1958; SEIBOLD & BERGER, 1996), or 271VaniËek, JuraËiÊ, BajraktareviÊ & ∆osoviÊ: Benthic Foraminiferal Assemblages in a Restricted Environment... STATIONS 1. Malo Jezero 2. Malo More 3. Veliko Jezero 4. Soline Channel 5. Kriæ (28 m) (13 m) (40 m) (1.5 m) (Open Sea) (38 m) salinity 36-38 34-36 35.5 -37 36-38 38 temperature 9-18 °C 9 -28 °C 9-17 °C 9-27 °C 13-16 °C oxygen concentration 0.0 -6.6 ml/l 4.8 -6.4 ml/l 0.0-6.6 ml/l 5 -6 ml/l 4.2 -6.6 ml/l Table 1 Ranges of bottom water characteristics at investigated stations. Summarized data after BULJAN & ©PAN (1976) and BENOVI∆ et al. (2000). Fig. 2 Schematic profile through the Mljet Lakes showing annual oscillation of temperature and the position of the thermocline, bottom sedi- ment type, and prevalent distribution of morphotypes of Elphidium crispum (A = morphotype #1; B = morphotype #2) and occurrence o f Elphidium aculeatum (C). Scale bar is 200 µm. 272 Geologia Croatica 53/2 lamination due to episodic bottom-water anoxic events which need not necessarily correspond to annual varia- tions (JURA»I∆ et al., 1995). 3. SAMPLING AND METHODS 3.1. SAMPLING AND SAMPLE PREPARATION Scuba divers collected sediment cores up to 80 cm long from five stations in the Mljet Lakes and the adja- cent open sea (Fig. 1) in May 1995. Sediment cores were frozen within 4 hours, and transported frozen to the laboratory. In the laboratory, frozen cores were cut into sub- samples, and the first 2-cm subsamples were used for our research. After defrosting, a subsample of 10 ml of wet sediment was taken for foraminiferal analysis, and the rest of the subsample was used for grain-size analy- sis. 3.2. METHODS The grain size of the sediment samples was ana- lyzed by wet sieving, using ASTM standard stainless steel sieves for >32 µm fraction. A Coulter Counter (mod. TA II) was used for the <32 µm sizes. The sedi- ments were classified according to their gravel - sand - mud (mud = silt + clay) ratio (FOLK, 1954). Carbonate content was determined volumetrically by measuring CO2 evolved by dissolving a 0.5 g dry sample in 15% HCl. The subsamples for foraminiferal analysis were stained with rose Bengal (MURRAY, 1973), in spite of the disadvantages and limitations of this method (MU- RRAY, 1991; JORRISEN et al., 1995). After staining, subsamples were wet sieved through 250 and 125 µm sieves. In each sample foraminifera were identified after FORAMINIFERI PADANI (1982), JORRISEN (1988), LOEBLICH & TAPPAN (1987), and CIMERMAN & LANGER (1991), and all specimens (stained and un- stained) were counted. The absolute abundance in 10 ml, and the relative abundance of each species was determined. For each station, dominant and accessory species were defined according to their relative abun- dance. An estimation of the species diversity was per- formed using the following parameters: number of species, number of specimens, Shannon-Wiener species diversity index (H), equitability index (E) (BUZAS, 1979), and percentage of Miliolina (=p o r c e l l a n e o u s ) , Textulariina (=agglutinated) and Rotaliina s . l . (=h y a- line). Species diversity and abundance were compared with the known ecological conditions (temperature, salinity, substrate). The morphotype variations were described from >125 µm size-group foraminifera. Existing corollary hydrographic data gave a better understanding about the ecological affinity of investi- gated species. Sedimentation rate was estimated accord- ing to the relative abundance of specimens of sessile species (POAG et al., 1980). 4. RESULTS 4.1. GRANULOMETRIC ANALYSES Results of granulometric analyses (mean size, sort- ing, mud percent, and sediment type), along with car- bonate content of investigated samples, are reported in Table 2. The sample from the deepest part of Malo Jezero (sample # 1; 28 m depth) has only 4.9% biogenous par- ticles/fragments of gravel and sand size, and the rest is silt and clay. The surface sediment is 71% carbonates, mostly aragonite, with lesser amounts of calcite, mag- nesian calcite, and dolomite. The remaining part is quartz and feldspar (SONDI et al., 1995). In the shal- low part of Malo More (sample # 2; 13 m depth), the WATER MEAN SORTING MUD SEDIMENT CARBONATE SAMPLE DEPTH SIZE, Mz So (%) TYPE (%) (m) (µm) (Folk, 1954) 1. Malo 1.61 slightly Jezero 28 10.9 poorly 95.1 gravelly 72 sorted mud 2. Malo 2.81 slightly More 13 176.8 very poorly 48.4 gravelly 89 sorted sandy mud 3. Veliko 3.45 gravelly Jezero 40 59.5 very poorly 73.3 mud 62 sorted 4. Soline 1.44 gravelly Channel 1.5 692.6 poorly 5.1 sand 76 sorted 5. Kriæ 2.02 gravelly (Open Sea) 38 1659.8 very poorly 12.2 muddy 59 sorted sand Table 2 Granulometric properties, types of sediments, and carbonate content. Surface sediment sample (0-2 cm), Mljet, May 1995. 273VaniËek, JuraËiÊ, BajraktareviÊ & ∆osoviÊ: Benthic Foraminiferal Assemblages in a Restricted Environment... sediment has 51.5% biogenous gravel and sand; the remainder is mud. Sample # 3 (40 m depth) in Veliko Jezero contains 26.5% biogenous particles (>125 µm). The bottom of the shallowest part of the Soline Channel (sample # 4; 1.5 m depth) contains carbonate gravely sand, which forms symmetrical ripples. Their wave lengths are 3 -6 m, as a consequence of the relatively strong tidal currents in that shallow channel. Sediment sorting is therefore better than in other samples (although still low =1.44). The nearshore marine sam- ple # 5 (Kriæ, 38 m depth) had the largest mean size (1660 µm), with 21.42% gravel, 66.43% sand, and the minimum carbonate. 4.2. DIVERSITY OF SPECIES AND ASSEMBLAGES, MORPHOTYPE VARIATIONS, AND SEDIMENTATION RATE The numbers of stained, unstained, and total speci- mens, along with the number of species present in each sample, are reported in Table 3. In total, 67 benthic foraminiferal species were determined, and 44 species were represented by stained specimens in at least one sample (Table 4). The ratio between unstained and stained specimens (U/S) varies from 10:1 (# 1 and 2) to 4:1 (# 4) (Table 3). Six species were found only in the marine lakes, and 10 were restricted to the open water. Dominant and accessory species were determined at each station (Table 5). In Table 6 the Shannon-Wiener index (H), equitability index (E), and percentage of Textulariina, Miliolina and Rotaliina are shown. In the sample from Malo Jezero (Station # 1; 28 m depth), we found 12 species, represented by only 132 specimens in the standard 10 ml sample. The U/S ratio is high (10:1). The majority of foraminifera belongs to the suborder Rotaliina (77% ). The specimens of Mili- olina are rare. This association is characterized by thin- walled species (sizes less than < 250 µm): H a y n e s i n a d e p r e s s u l a , Asterigerinata mamilla , Valvulineria bra - dyana. H index is the lowest of all samples (0.81), while the E index is 0.22. We refer to this association as the Haynesina depressula Assemblage (Table 5). At 40 m depth in Veliko Jezero (Station # 3), we found the Asterigerinata mamilla Assemblage (Table 5). This sample contains 18 species and 146 specimens, almost entirely belonging to the Rotaliina (91 % ) : A . m a m i l l a, V. bradyana , H. depressula , Rosalina bradyi and a large Elphidium crispum (Tables 3, 4). U/S ratio is low (5 :1), and H and E indices are low (0.97 and 0.23 respecively). In Malo More and in the Soline Channel we found a foraminiferal fauna different from that in Malo Jezero and Veliko Jezero. In Malo More (Station # 2; 13 m depth), 32 species and 978 specimens were determined (U/S = 10:1, H= 0.91, E = 0.18), and in Soline Channel (Station # 4; 1.5 m depth), 38 species with 1810 speci- mens were counted (U/S = 4:1, H =1.23, E= 0.23) (Tab- les 3, 6). We assigned benthic foraminiferal assem- blages in Malo More to the Peneroplis planatus Assem- blage, whereas the Peneroplis pertusus A s s e m b l a g e characterizes the station in Soline Channel (Table 5). In Malo More, P. pertusus and P. planatus p r e v a i l and constitute about 60% of the foraminiferal assem- blage, along with small Rotaliina: epifaunal A. mamilla and infaunal H. depressula . In the Soline Channel, the same species as in the “lakes” were found (E. crispum, A. mamilla, H. depressula ), along with sessile and attached species (the later are sometimes mobile): Cibi - cides refulgens, Cibicidella variabilis , Lobatula lobatu - l a , Planorbulina mediterranensis , R. bradyi, G a v e l i - nopsis lobatulus. The open sea Kriæ sample (Station # 5; 38 m depth) was the only one in which T e x t u l a r i i n a exceeded 1% . The Rosalina bradyi - Cibicides refulgens A s s e m b l a g e contained 55 species with 876 specimens (H = 1 . 4 8 , E= 0.26) (Tables 3, 5, 6). Biogenous fragments form an ideal substrate for such sessile species. Beside R. bradyi and C. refulgens, the sample contained G. lobatulus , A. NUMBER OF DRY SPECIMENS UNSTAINED STAINED Ratio TOTAL SAMPLE WEIGHT IN GRAM OF SPECIES SPECIMENS SPECIMENS U/S SPECIMENS IN 10 ml (g) SEDIMENT (U) (S) 1. Malo Jezero 8.29 15.9 12 120 12 10:1 132 2. Malo More 9.59 102.0 32 887 91 10:1 978 3. Veliko Jezero 6.29 23.2 18 122 24 5:1 146 4. Soline Channel 9.23 196.1 38 1462 348 4:1 1810 5. Kriæ (Open Sea) 10.02 87.4 55 771 105 7:1 876 Table 3 Number of species, unstained and stained individuals and total number of specimens in each sample, Mljet, May 1995. 274 Geologia Croatica 53/2 STATIONS 1. MALO 2. MALO 3. VELIKO 4. SOLINE 5. KRIÆ JEZERO MORE JEZERO CHANNEL (OPEN SEA) unstained stained unstained stained unstained stained unstained stained unstained stained FORAMINIFERA No. % No. % No. % No. % No. % No. % No. % No. % No. % No. % Total 120 100 12 100 887 100 91 100 122 100 24 100 1462 100 348 100 771 100 105 100 TEXTULARIINA Textularia sp. 1 0.11 1 0.9 11 1.4 MILIOLINA Vertebralina striata 21 2.4 1 1.1 8 0.6 1 3 3 0.4 Adelosina mediterranensis 3 0.3 10 1.3 Spiroloculina ornata 7 0.8 2 2.2 2 0.1 1 0.3 14 1.8 3 2.9 Siphonaperta aspera 9 1.0 72 4.9 18 5.2 Cycloforina contorta 3 0.2 5 0.7 C. juleana 10 0.7 4 0.5 C. villafranca 3 0.4 2 1.9 C. colomi 1 0.1 Lachlanella variolata 1 0.1 10 0.7 4 1.2 1 0.1 Massilina secans 1 0.1 1 1.0 M. gualtieriana 1 0.1 Quinqueloculina nodulosa 4 0.5 2 0.1 23 3.0 Q. laevigata 18 15 3 25 29 3.3 8 8.8 3 2.5 11 0.8 9 1.2 1 1.0 Q. limbata 3 4 Q. stelligera 8 1.0 Q. jugosa 2 0.6 24 3.1 Q. parvula 3 2.5 25 2.8 188 13.0 9 2.6 17 2.2 3 2.9 Q. bidentata 36 2.5 6 1.7 8 1.0 2 2.0 Q. seminula 52 3.6 8 2.3 3 0.4 2 2.0 Q. ungeriana 6 0.4 3 0.9 6 0.8 1 1.0 Miliolinella subrotunda 2 0.3 3 2.9 M. semicostata 2 0.2 3 0.4 1 1.0 M. grata 2 0.1 4 1.2 M. webbiana 1 0.1 Pseudotriloculina laevigata 1 0.1 10 0.7 17 2.2 1 1.0 P. oblonga 4 3.3 2 0.2 2 1.6 8 0.6 4 0.5 1 1.0 Triloculina marioni 1 0.1 1 1.1 10 0.7 10 1.3 1 1.0 T. schreiberiana 1 0.1 41 3.0 9 2.6 13 1.7 1 1.0 Pyrgo sp. 2 0.26 Sigmoilinita costata 1 0.1 34 4.41 Articulina carinata 1 0.13 Coscinospira hemprichii 1 0.1 Wellmanelinella striata 2 0.1 Laevipeneroplis inornatus 1 0.1 Peneroplis pertusus 2 1.7 231 26.0 26 28.6 2 1.6 1 4.17 309 21.1 70 20.1 10 1.3 2 1.9 P. planatus 287 32.4 30 33 1 0.8 1 4.17 288 19.7 54 15.5 7 0.91 P. arietinus 4 0.3 Sorites orbiculus 2 1.6 ROTALIINA Lenticulina sp. 3 0.4 Polymorphina sp.7 8 1.0 1 1.0 Sphaerogypsina globula 9 1.2 Bolivina sp. 6 5 1 8.3 1 0.1 1 0.8 Elphidium crispum 4 3.3 22 2.5 1 1.1 18 14.8 7 29.2 93 6.4 80 23.0 46 6.0 10 9.5 E. aculeatum 3 0.2 1 0.3 16 2.1 E. sp. 1 1 0.1 1 1.1 3 2.5 10 1.3 E. sp. 4 2 1.7 3 0.3 2 1.6 29 2.0 5 1.4 32 4.2 1 1.0 E. cf. depressulum 4 0. 5 38 2.6 9 2.6 E. cf. maioricensis 1 0.1 7 0.9 Planorbulina mediterranensis 1 0.1 6 0.4 3 0.9 6 0. 8 3 2.9 Cibicides refulgens 1 0.1 2 2.2 2 1.6 2 8.3 45 3.1 12 3.6 77 10.0 6 5.7 Lobatula lobatula 3 0.3 3 2.5 29 2.0 3 0.9 13 1.7 5 4.7 Cibicidella variabilis 3 0.2 3 0.9 1 1.0 Asterigerinata mamilla 3 2.5 107 12.1 10 11 37 30.3 9 37.5 6 0.4 2 0.6 67 8.7 7 6. 7 Gavelinopsis lobatulus 3 2.5 93 6.4 31 8.9 42 5. 5 13 12.4 Rosalina bradyi 1 0.8 26 2.9 1 1.1 7 5.7 32 2.2 6 1.7 73 9.5 18 17.1 R. macropora 3 0.4 3 2.9 R. vilardeboana 27 3.5 2 1.9 Haynesina depressula 39 32.5 8 66.7 36 4.1 8 8 14 11.5 3 12.5 1 0.3 Valvulineria bradyana 20 16.7 7 0.8 21 17.2 1 4.2 2 1.9 Eponides concameratus 1 0.1 Neoconorbina terquemi 48 6.2 5 4.8 Astronion stelligerum 2 0.1 13 1.7 3 2.9 Ammonia parkinsoniana 6 0.4 3 0.9 A. tepida 18 15 48 5.4 2 0.3 Lagena sp. 2 0.3 Buccella sp. 1 8 1.1 275VaniËek, JuraËiÊ, BajraktareviÊ & ∆osoviÊ: Benthic Foraminiferal Assemblages in a Restricted Environment... mamilla, E. crispum and in much smaller amount spine specimens of E. aculeatum. Several different species of Miliolina were also present (Table 4). Morphotype variations were only found in the Elphidium crispum species (Fig. 2). Two morphotypes have been distinguished based on the following features (after JORRISEN, 1988): (1) maximum diameter; (2) number of chambers per final whorl; (3) structure and ornamentation of the umbilical part of the tests (the presence or absence of defined umbilical boss); (4) degree of inflation of the chambers (their shape); (5) outline of the test. Morphotype # 1 of E. crispum species is a thin-shelled, flattened form with the maxi- mum size reaching 1.44 mm, with a high number of arcuate chambers (27-29) in the last whorl, and with 12 - 14 ponticuli with subelliptical fossettes in last chamber. The sutures are backward curved, with a pro- nounced umbilical knob with 7 -9 large perforations, and with angular peripheral margin. Morphotype # 2 is characterized by thick-shelled, lenticular, inflated tests, the maximum size is up to 0.57 mm; low number of arcuate chambers (10 -14) in the last whorl, 9 -12 ponticuli in the last chamber, the fos- settes between ponticuli are rectangular in shape, cham- bers gradually increasing in size. The sutures are back- ward curved, with a flat umbilical knob. In Malo More and Veliko Jezero (samples # 2 and # 3 respectively), morphotype # 1 prevails (in sample # 2 morphotype # 1 makes 90% of all elphidiids found at this station, and in sample # 3 it participates with 67%). In sediment samples from the Soline Channel and Kriæ (# 4 and 5), morphotype # 2 strongly prevails (87% and 91% respectively) (Fig. 2). Along with the distribution of morphotypes of E . c r i s p u m, the occurrence of E. aculeatum is shown in Fig. 2. It was found only in samples from Stations # 4 , and # 5. Bottom sediment type at investigated locations is also presented in Fig. 2. The percentage of specimens of attached species C i b i c i d e s, Rosalina and Planorbulina differs signifi- cantly from station to station (Fig. 3). In Malo Jezero, specimens of these species were not recorded, but toward the open sea their percentage significantly increases, up to 11.7% in the unstained assemblage and 1 4 .7% in the stained assemblage at the station Kriæ. The percentage of sessile specimens in the stained Table 4 Benthic foraminiferal abundance. Surface sediment sample (0-2 cm), Mljet, May 1995 (numbers of stained individuals are given in the right column and numbers of unstained are given in the left column for each station). UNSTAINED ASSEMBLAGES STAINED ASSEMBLAGES STATIONS dominant secondary dominant secondary species species species species 1. Malo Jezero Haynesina Valvulineria Haynesina Quinqueloculina depressula bradyana, depressula laevigata Ammonia tepida, Quinqueloculina laevigata 2. Malo More Peneroplis Peneroplis Peneroplis Peneroplis planatus pertusus, planatus pertusus, Asterigerinata Asterigerinata mamilla mamilla 3. Veliko Jezero Asterigerinata Valvulineria Asterigerinata Elphidium mamilla bradyana, mamilla crispum Elphidium crispum, Haynesina depressula 4. Soline Channel Peneroplis Quinqueloculina Elphidium Peneroplis pertusus, parvula crispum pertusus, Peneroplis Peneroplis planatus planatus 5. Kriæ (Open Sea) Cibicides Asterigerinata Rosalina Elphidium refulgens mamilla, bradyi crispum, Gavelinopsis Gavelinopsis lobatulus lobatulus Table 5 Dominant and secondary species of benthic foraminifera in samples from each station. Surface sediment sample (0-2 cm), Mljet, May 1995. 276 Geologia Croatica 53/2 assemblage is larger than the percentage of sessile spec- imens in the unstained association, contrary to the majority of vagile species. 5. DISCUSSION As expected from physiographic and hydrographic data, the benthic foraminiferal assemblages show large differences. The ratio of Miliolina, Textulariina and Rotaliina (Table 6) indicates that the assemblages from stations located in the “lakes” (Stations # 1 and 3) dis- play characteristics of lagoon assemblages (MURRAY, 1991), whereas the assemblage at the marine nearshore station Kriæ (Station # 5) indicates a normal marine environment with salinity over 30. The smallest number of species and specimens in a standard 10 ml sample was found in Malo and Veliko Jezero (samples # 1 and 3). Such assemblages are oligospecific (showing low species diversity and low abundance), but at the same time have high species dominance. Generally, they indicate a specific, restrict- ed environment with stressed conditions (oxygen deple- tion, unusual temperatures, low or high salinity or their large variations). SEN GUPTA & MACHAIN-CASTI- LLO (1993) found that in dysoxic conditions typically 2-3 species constitute up to 80% of the total assem- blage. We infer that the primary limiting factor for low species abundance, low species diversity, and high dominance in the Mljet Lakes is episodic low oxygen concentration (Table 1). The oxygen depletion of bot- tom water is produced by density stratification; in Malo Jezero the thermocline occurs between 5 and 12 m depth and in Veliko Jezero between 12 and 22 m depth. Both samples show that the fauna have three species making up to 64% of the total (unstained and stained) assemblage; H. depressula , V. bradyana and A. tepida represent 64% of total assemblage in sample # 1; A . m a m i l l a, V. bradyana a n d E. crispum make 63% of total assemblage in sample # 3. In these samples the second dominant species is V. bradyana (making up to 15% of total assemblage at each station), an oxygen deficiency tolerant species (JORISSEN, 1987, p. 32). This species is absent in samples # 4 and 5. The species diversity index H for total (stained and unstained) assemblages increases from about 0.8 in sample # 1 to 1.4 in sample # 5. Lower values found in samples # 1, 2 and 3 are the result of a high dominance of very few species (Table 5). The highest value (sam- ple # 5) implies that more favourable conditions for foraminiferal diversity exist in the open marine envi- ronment. The equitability index E is very low ranging from 0.18 in sample # 2 to 0.26 in sample # 5. This implies that all the samples are from somehow restrict- ed environments. Reoxygenation of bottom water after episodic defi- ciencies in the Mljet Lakes, as in other environments, is followed by colonization of benthic foraminifera. Shal- low water foraminifera require only about three weeks to colonize the substrate and to stabilize their density (BUZAS, 1993, p. 158). Amongst many genera, repre- sentatives of Q u i n q u e l o c u l i n a and E l p h i d i u m are con- sidered to be colonization pioneers. Their presence in all samples does not allow determination of whether foraminifera lived through the low oxygen/anoxia events, or if they recolonized after. However, in the Malo Jezero sample # 1, the ratio between quinquelo- culinids and elphidiids (0.18 :0.05) is similar to that known from an assemblage found at 1 m depth in experimental conditions in Florida (0.35 :0.07 - BUZAS, 1993, p. 159). The different ratio in the Veliko Jezero sample # 3 (0.02:0.21) probably reflects some- what different conditions (greater depth, diminished turbulence, no algal cover, lower temperature) which are favorable for elphidiids. The tendency of changing shell shape with depth has been observed as intraspecific variation in operculi- nas (PECHEUX, 1995). The morphological changes have been described either as thinning of the walls, or as a change in chamber proportions (HOTTINGER, 1997). In lamellar-perforate foraminifera the thinning of the chamber walls has been shown to be the result of diminishing rates of turbulence in the water column (REISS & HOTTINGER, 1984). The distribution of the morphotypes of E. crispum in the Mljet Lakes coincides with differences in the nature of substrate. Muddy sediments with a high food concentration (samples # 2, 3) host prevalently large, flattened, morphotype # 1, whereas in a sandy substrate (samples # 4, 5) smaller, inflated, morphotype # 2 (with diameters about 40 % smaller than those of morpho- SHANNON- EQUITA- PERCENTAGE OF SAMPLE WIENER BILITY TEXTULARIINA MILIOLINA ROTALIINA INDEX INDEX NUMBER % NUMBER % NUMBER % (H) (E) 1. Malo Jezero 0.81 0.22 0 0 30 22.73 102 77.27 2. Malo More 0.91 0.18 1 0.1 694 71.27 283 28.94 3. Veliko Jezero 0.97 0.23 1 0.68 12 8.22 133 91.1 4. Soline Channel 1.12 0.23 0 0 1265 69.89 545 30.11 5. Kriæ (Open Sea) 1.48 0.26 11 1.26 272 31.05 593 67.7 Table 6 Shannon-Wiener index, equitability index, and percentage of Textulariina, Miliolina and Rotaliina, Mljet, May 1995. 277VaniËek, JuraËiÊ, BajraktareviÊ & ∆osoviÊ: Benthic Foraminiferal Assemblages in a Restricted Environment... types # 1) prevails. This contrasts JORISSEN’s (1988) finding from the northern Adriatic Sea, where large, compact, morphotypes prevail in nutrient-poor condi- tions, and inflated morphotype in nutrient-rich condi- tions. On the other hand, this goes along with experi- mental evidence (TER KUILE & EREZ, 1988) that in lamellar perforate foraminifera the thinning of the chamber walls is a result of thinning of each lamella due to the diminishing rates of turbulence with increase in depth. MURRAY (1991) suggested that elphidiids from a muddy substrate have infaunal preferences, while those from sand are more likely to have an epifaunal prefer- ence. LANGER (1988) and KITAZATO (1988) have described an epiphytic mode of life for E. crispum specimens, while MYERS (1943) noted that larger E . c r i s p u m, buried to a depth of about 1 cm, are able to escape in about 1 hour (calculation is made after Myers data, where active individuals spend about 12 hours tra- versing distance of 9 to 12 cm in vertical direction). It is expected that interstitial water in subsurface sediment (at one cm depth) has lower oxygen concentration than water at the water/sediment interface. Therefore, thin- ner, larger specimens (morphotype # 1) prefer muddy sediment with calm water above (Table 1; Fig. 2). The relative sediment accumulation (relative sedi- mentation rate) can be estimated by using the percent- age of sessile species, because attached species are incompatible with large sediment input (POAG et al., 1980). There is a net increase of the percent of speci- mens of sessile species from Malo Jezero (absent) toward the open sea (Station # 5, Kriæ) (Fig. 3). It sug- gests a greater sedimentation rate in the “lakes” than in open waters. The “lakes” most probably act as a trap for terrigenous and authigenous particles due to their topography. Haynesina depressula is one of the most abundant species in bottom sediments of both Malo Jezero and Veliko Jezero. MURRAY (1991, p. 324), in an ecologi- cal overview of foraminifera, has representatives of genus H a y n e s i n a as brackish forams (living in salinity between 0 and 30), although ( ibid, p. 141) some species can survive higher salinity (0-35). At the Malo Jezero, Malo More and Veliko Jezero stations stained speci- mens were noticed, indicating that those specimens were most probably authigenic. At the open-sea Kriæ station, H. depressula was not recorded, although hydrographic conditions, except for oxygen concentra- tion, are similar to those in the “lakes”. The percentage of H. depressula , similarly to V. bradyana, in the asso- ciation decreases towards the open sea, whereas at the same time the biodiversity increases. Most probably, H. d e p r e s s u l a tolerates stress conditions including a large range of oxygen concentrations, salinity and tempera- ture, but is probably a bad competitor. The lack of H . depressula at the Kriæ station might therefore be a con- sequence of the increased competition there. 6. CONCLUSIONS This study is based on five surface sediment sam- ples collected in May, 1995, from the Mljet Lakes, and therefore can be considered only as a preliminary note. However, it is possible to put forward the following conclusions: (1) Benthic foraminiferal assemblages from the Lakes area and from the nearshore indicate a transition from lagoonal towards open marine conditions, based on the ratio between Textulariina, Miliolina and Rotaliina. (2) The “lakes” behave as temporary hypoxic lagoons, in which the main limiting factor for the benthic foraminifera is oxygen concentration. Episodic stag- nant conditions in Malo and Veliko Jezero are Fig. 3 Proportion of sessile spe- cimens of C i b i c i d e s, P l a - norbulina , Rosalina in inve- stigated samples. 278 Geologia Croatica 53/2 formed due to large temperature and salinity differ- ences between bottom and surface water, which pre- vents water mixing. At the same time the “lakes” act as sediment traps with a higher sedimentation rate compared to the adjacent open sea. Episodic anoxia is reflected in a low foriminiferal species diversity, high abundance of specimens of two pioneer genera (Quinqueloculina a n d E l p h i d i u m), and presence of V a l v u l i n e r i a species tolerant to low oxygen condi- tions. (3) The number of species and specimens in sediment foraminiferal assemblages increases with the influ- ence of the open sea. In these peculiar marine lakes, unstable ecological conditions do not permit the accommodation of either a large number of species or of specimens, whereas in the nearshore, where ecological conditions are more stable (“normal”), the assemblage is much richer in species. The Shan- non-Wiener index is rather low in all samples sug- gesting restricted environment (H = 0 .8 -1 . 4 ) , although its value increases continuously from the Malo Jezero towards the Kriæ Station. (4) The distribution of morphotypes of E l p h i d i u m c r i s p u m in the Mljet Lakes may be correlated to substrate type, mode of life, water energy, and oxy- gen conditions. The larger flattened specimens of morphotype # 1 occur more commonly in a muddy, oxygen depleted, and low energy environment. (5) In the study area, Haynesina depressula tolerates a salinity range significantly higher than previously documented. It appears to be well adapted to stressed conditions, but probably cannot tolerate competition. Acknowledgments The authors wish to thank the Mljet National Park authorities for financial support for field research, Dr. Ivan SONDI for help in sampling, Miss Renata SLAV- KOVI∆ (INA - Naftaplin, Zagreb), Mr. Momir MILU- NOVI∆ (Ruer BoπkoviÊ Institute, Zagreb) and Mr. Miroslav KLADNI»KI (Institute of Geology, Zagreb) for technical assistance. We also thank to the referees Prof. dr. Ivan GU©I∆ and Dr. Franc CIMERMAN for their helpful comments. 7. REFERENCES B E N O V I ∆ , A. & O N O F R I , V. (1995): Ekoloπke oso- bitosti Velikog jezera na otoku Mljetu.- In: DUR- BE©I∆, P. & BENOVI∆, A. (eds.): Prirodne znaËaj- ke i druπtvena valorizacija otoka Mljeta. Conf. proc. Pomena, otok Mljet, September 4-10, 1995, Hrvat- sko ekoloπko druπtvo, Ekoloπke monografije, 6, Zagreb, 511-521. BENOVI∆, A., LU»I∆, D., ONOFRI, V., PEHARDA, M., CARI∆, M., JASPRICA, N. & BOBANOVI∆- »OLI∆, S. (2000): Ecological characteristics of the Mljet Island seawater lakes (South Adriatic Sea) with special reference to their resident populations of medusae.- Scientia Marina (in press). B U L J A N, M. & © P A N, J. (1976): Hidrografska svojst- va mljetskih jezera i susjednog mora.- Acta Adriati- ca, 6/12, 1-227. BULJAN, M. & ZORE-ARMANDA, M. (1976): Ocea- nographic properties of the Adriatic Sea.- Oceanogr. Mar. Biol. Ann. Rev., 14, 11-98. B U Z A S, M.A. (1979): The measurement of species diversity.- In: LIPPS, J.H., BERGER, W.H., BUZ- AS, M.A., DOUGLAS, R.G., & ROSS, C.A. (eds.): Foraminiferal Ecology and Paleoecology. Soc. Econ. Paleont. Mineral., Short Course, 6, 3-10. B U Z A S, M.A. (1993): Colonization rate of foramini- fera in the Indian River, Florida.- Jour. of Foram. Res., 23/3, 156-161. C A R I ∆, M. & J A S P R I C A, N. (1995): Koncentracija hranjivih soli i klorofila α u Velikom jezeru u raz- doblju stratifikacije.- In: DURBE©I∆, P. & BEN- OVI∆, A. (eds.): Prirodne znaËajke i druπtvena val- orizacija otoka Mljeta. Conf. proc. Pomena, otok Mljet, September 4-10, 1995, Hrvatsko ekoloπko druπtvo, Ekoloπke monografije, 6, Zagreb, 447-451. C I M E R M A N, F., D R O B N E, K. & O G O R E L E C, B. (1988): L’association de foraminiferes benthiques des vases de la baise de Veliko Jezero sur l’île de Mljet et de la falaise Lenga, Ouverte vers la mér (Adriatique moyenne).- Revue de Paléobiologie, Vol. spéc. 2 (Benthos ′86), 741-753. C I M E R M A N, F. & L A N G E R, M. R. (1991): Mediter- ranean Foraminifera.- Dela SAZU IV razreda, 1-95, pls. 1-118, Ljubljana. F O L K, R.L. (1954): The distinction between grain size and mineral composition in sedimentary-rock nom- enclature.- Jour. of Geology, 62, 344-359. FORAMINIFERI PADANI (1982): Atlante iconografi- co e distribuzione stratigrafica (Terziario e Quater- nario).- Sec. edizione, Agip S.p.A. Milano, 52 pls. H O T T I N G E R, L. (1997): Shallow benthic foraminifer- al assemblages as signals for depth of their deposi- tion and their limitations.- Bull. de la Soc. Geol. de France, 168/4, 491-505. J O R I S S E N, F.J. (1987): The distribution of benthic foraminifera in the Adriatic Sea.- Marine Micropal., 12/1, 21-48. J O R I S S E N, F.J. (1988): Benthic foraminifera from the Adriatic Sea: principles of phenotypic variation.- Utrecht Micropal. Bull., 37, 176. J O R I S S E N, F.J., DE STIGTER, H.C. & W I D M A R K, J.G.V. (1995): A conceptual model explaining ben- 279VaniËek, JuraËiÊ, BajraktareviÊ & ∆osoviÊ: Benthic Foraminiferal Assemblages in a Restricted Environment... thic foraminiferal microhabitats.- Marine Micropal., 26, 3-15. JURA»I∆, M., SONDI, I., BARI©I∆, D., VDOVI∆, N. & PR A V D I ∆, V. (1995): Sedimenti i sedimentacija u Mljetskim jezerima.- In: DURBE©I∆, P. & BEN- OVI∆, A. (eds.): Prirodne znaËajke i druπtvena val- orizacija otoka Mljeta. Conf. proc. Pomena, otok Mljet, September 4-10, 1995, Hrvatsko ekoloπko druπtvo, Ekoloπke monografije, 6, Zagreb, 107-116. K I T A Z A T O, H. (1988): Ecology of benthic forami- nifera in the tidal zone of a rocky shore.- Revue de Paléobiologie, Vol. spec. 2 (Benthos ‘86), 815-825. L A N G E R, M. (1988): Recent ephiphytic foraminifera from Vulcano (Mediterranean Sea).- Revue de Paléobiologie, Vol. spec. 2 (Benthos ‘86), 827-832. L O E B L I C H, A.R. & T A P P A N, H. (1987): Foramini- feral genera and their classification.- Van Nostrand Reinhold, New York , Vol. I X+970 p., Vol. II VIII+212 p. (pls. 1-847). M U R R A Y, J.W. (1973): Distribution and ecology of living benthic foraminiferids.- Heinemann, London, 294 p. M U R R A Y, J.W. (1991): Ecology and paleoecology of benthic foraminiferes.- Longman Scientific and Te- chnical, Harlow, Essex, 391 p. MYERS, E.H. (1943): Life activities of foraminifera in relation to marine ecology.- Proc. of the Amer. Phil. Soc., 86/3, 439-457. OPEN UNIVERSITY COURSE TEAM (1989): Sea- water: its composition, properties and behaviour.- Pergamon Press, Oxford, 165 p. PECHEUX, M.J-F. (1995): Ecomorpholoy of a recent larger foraminifer Operculina ammonoides.- G e o- bios, 28/5, 529-566. POAG, C.W., KNEBEL, J.H. & T O D D, R. (1980): Distribution of modern benthic foraminifers on the New Jersey outer continental shelf.- Marine Micropal., 5/1, 43-69. REISS, Z. & HOTTINGER, L. (1984): The Gulf of Aqaba, ecological micropaleontology.- Ecological studies 50, Springer Verlag, Berlin, 354 p. S E I B O L D, E. (1958): Jahreslagen in sedimenten der mittleren Adria.- Geol. Rundschau, 47, 100-117. S E I B O L D, E. & B E R G E R, W.H. (1996): The Sea floor. An introduction to marine geology.- Springer Verlag, Berlin, 356 p. SEN GUPTA, B.K. & M A C H A I N - C A S T I L L O , M . L . (1993): Benthic foraminifera in oxygen-poor habi- tats.- Marine Micropal., 20, 183-201. S O K A », A. & B A J R A K T A R E V I ∆, Z. (1995): Mikro- organizmi i nanoorganizmi u sedimentima Malog i Velikog jezera otoka Mljeta.- In: DURBE©I∆, P. & BENOVI∆, A. (eds.): Prirodne znaËajke i druπtvena valorizacija otoka Mljeta. Conf. proc. Pomena, otok Mljet, September 4-10, 1995, Hrvatsko ekoloπko druπtvo, Ekoloπke monografij, 6, Zagreb, 125-146. SONDI, I., JURA»I∆, M., SLAVKOVI∆, R. & PRAV- D I ∆, V. (1995): Mineraloπke i povrπinske fizikalno kemijske znaËajke sedimenata mljetskih jezera.- In: P. DURBE©I∆, P. & BENOVI∆, A. (eds.): Prirodne znaËajke i druπtvena valorizacija otoka Mljeta. Conf. proc. Pomena, otok Mljet, September 4-10, 1995, Hrvatsko ekoloπko druπtvo, Ekoloπke mono- grafije, 6, Zagreb, 117-125. STRAÆI»I∆, N. (1979): Ljepote i znamenitosti Mljet- skog jezera.- Priroda, 1979/5, 144-150, Zagreb. TER KUILE, B. & EREZ, J. (1988): The size and func- tion of the internal inorganic carbon pool of the foraminifer Amphistegina lobifera.- Mar. Biol., 99, 481-487. VU»ETI∆, M. & VU»ETI∆, V. (1995): Klimatske pri- like otoka Mljeta kao Ëimbenik prilagodbe biljnog svijeta.- In: DURBE©I∆, P. & BENOVI∆, A. (eds.): Prirodne znaËajke i druπtvena valorizacija otoka Mljeta. Conf. proc. Pomena, otok Mljet, September 4-10, 1995, Hrvatsko ekoloπko druπtvo, Ekoloπke monografije, 6, Zagreb, 233-243. VULETI∆, A. (1953): Structure géologique du fond du Malo Jezero et du Veliko Jezero, sur l’île de Mljet.- Acta Adriatica, 6/1, 1-65. Manuscript received June 6, 2000. Revised manuscript accepted November 13, 2000. 280 Geologia Croatica 53/2