2016 | 69/2 | 205–212 | 2 Figs. | 5 Tabs. | www.geologia-croatica Journal of the Croatian Geological Survey and the Croatian Geological Society Traces of drilling predation in the Upper Badenian (Middle Miocene) molluscs from the Rakovica stream (Belgrade) Meri Ganić1, Predrag Radović2, Ljupko Rundić1, Katarina Bradić2 and Slobodan Knežević1 1 University of Belgrade, Faculty of Mining and Geology, Department of Regional Geology, Belgrade, Kamenička 6, Serbia; (corresponding author: meri.ganic@rgf.bg.ac.rs, ljupko.rundic@rgf.bg.ac.rs, slobodan.knezevic@rgf.bg.ac.rs) 2 National Museum Kraljevo, Trg Svetog Save 2, Serbia; (pedja_radovic@yahoo.com) 3 University of Belgrade, Faculty of Mining and Geology, Department of Paleontology, Belgrade, Kamenička 6, Serbia; (bradic.kaca@gmail.com) doi: 10.4154/gc.2016.14 Abstract The palaeoecology of fossil molluscs of Serbia is insufficiently studied. Drilling predation indi- cates interactive relationships among Middle Miocene Badenian molluscs from the Rakovica sands. These sands are present in the Rakovica stream valley (southern part of Belgra de), and well-known data about the fauna has existed since the end of the XIX century. The collected material (one bulk sample) from the sec tion was used for studies that included 2301 shells of molluscs (1617 of gastropods and 684 of bivalves). Among them were specimens with preda- tory drill holes. Determining the value of drilling frequency (DF) and prey effectiveness (PE) fa- cilitates identification of the rela tionship between shell-drilling predators and their prey. Based on analysis of the overall association, carnivorous gastropods were do mi nant in the sample, comprising more than fifty percent (52.32%) of all gastropod specimens. In addition, they con- stitute a large percentage of the total sample (43.19%). The largest number of victims was ob- served among bivalves, which are indicated by a DF of 16.96%. The results were compared with other localities of Miocene age in Paratethys and with studies on the Miocene sediments from the Boreal and the Atlantic Provinces. These comparisons show that PE and DF are very similar to results from the Central Paratethys, but that they are significantly higher than in the other are as mentioned. 1. INTRODUCTION Traces of biotic drillings found on skeletons of marine inverte- brates provide arguably the richest source of quantifiable data on prey-predator interactions available in the fossil record (KO WA- LEWSKI, 2002). Indeed, many studies of biotic interactions in the geological past have focused on drilling predation (VERMEIJ, 1982, 1987; VERMEIJ & DUDLEY, 1982; BAUMILLER, 1996; HOFFMEISTER & KOWALEWSKI, 2001; SAWYER & ZUS- CHIN, 2011; CHATTOPADHYAY & DUTTA, 2013). The over- all patterns of drilling predation are well established on Ce no zoic molluscs from the Atlantic coast of North America (KELLEY & HANSEN, 1993, 1996, 2006), but there has been considerably less works conducted in Europe. Relevant studies of drilling pre- dation on molluscs from the Miocene of Central Paratethys (HOFFMEISTER & KOWALEWSKI, 2001; SAWYER & ZUS- CHIN, 2011) did not include any data from Serbia, so details of predator-prey interactions in the marine Middle Miocene from this part of the southeastern margin of the Pannonian basin re- mained largely unknown. The Rakovica stream („Rakovički Potok”) near Belgrade re­ presents one of the richest Middle Miocene (Badenian) fossili- ferous sites in Serbia and has been the subject of geological and palaeontological studies for more than a century. PAVLOVIĆ (1890, 1893, 1922, 1925) and PAVLOVIĆ & PETKOVIĆ (1903) studied the fossil fauna itself, while LUKOVIĆ (1922), STE VA­ NO VIĆ & STEPANOVIĆ (1939), STEVANOVIĆ (1951) and E RE MIJA (1977, 1987) focused mainly on the stratigraphy of the site in the regional comparative context. PAVLOVIĆ (1925) was the first to note traces of biotic drillings on fossil specimens from the Rakovica stream. He discovered that drilled shells belonged to ten gastropod and four bivalve species, and he assumed that these biotic boreholes were produced by predatory snails (Natica and Conus) found among other taxa on the site. Later study by KOHN (1983) reported that drilling predation is characteristic for Naticidae and Muricidae. Similar palaeoecological observations were discovered by STEVANOVIĆ (1970), who de scribed traces of predation on Badenian shells collected in the vicinity of Bel- grade, and more recently by JOVANOVIĆ (2002), who examined a Badenian fauna from Golubac, north-eastern Serbia. However, none of these studies focused on drilling predation and they were not carried out in a systematic manner. In order to fill this gap in knowledge, we present here the first detailed report of predatory behaviour within molluscs from the Middle Miocene of Serbia, which includes the comparison with the available data for Para- tethys, and the Boreal and North Atlan tic Provinces (HOFF- MEISTER & KOWALEWSKI, 2001, KELLEY & HANSEN, 2006, SAWYER & ZUSCHIN, 2011). 2. GEOLOGIC SETTING The fossil sample used for the study was collected from the Ra- kovica stream section, („Rakovički Potok”) located on the left bank of Rakovica creek in the southern part of Belgrade (WGS 84, 44°43′58″, 20°30′27″). In Serbia, the site represents a classic locality of the sandy facies of the Badenian (Middle Miocene), the so-called Rakovica sands which are rich in marine fossils. It is located on the Basic geological map at scale 1:100000, sheet Pančevo (IVKOVIĆ, 1966). This area, based on main stratigra- phic and tectonic studies, is a composite geological setting rep- resented by block (horst and graben) structures. In this relatively Article history: Manuscript received August 06, 2015 Revised manuscript accepted January 27, 2016 Available online May 16, 2016 Keywords: Gastropods, Bivalves, Statistical analysis, Predator-prey interactions, Palaeoecological relations G eo lo gi a C ro at ic a Geologia Croatica 69/2206 small area and during the younger phases of Neoalpine tectonics there were significant different movements which formed a mo­ dern relief. The Torlak hill is a horst structure with Mesozoic rocks in its core part, while the slopes are composed of different Ba- denian and Sarmatian deposits (MAROVIĆ et al., 2007). Further downstream from the Avala road, in the middle part of Rakovica stream, there is a small graben structure which is part of the Beli Potok graben, which is filled with the upper Miocene sediments (Pannonian and Pontian) as well as those of middle Miocene (Sar- matian) age. The Torlak horst and the Beli Potok graben are se- parated by the so-called Rakovica fault. These geological struc- tures are characterized by large differential movements along the fault with a vertical displacement of more than 200 m. For ex- ample, in borehole KGK-14 below the Pontian sediments, at an altitude of 100 m, the Pannonian marls were discovered (KNE- ŽE VIĆ, 1989). However, just a few hundred metres away to north-east, in the Torlak foothill, the same Pannonian marls oc- cur on the surface near 210 m. Similarly, at the top of Torlak hill, Badenian sediments are present at the surface at 336 m. All tec- tonic movements took place during the late Miocene and Pliocene as well as the early Pleistocene. Based on strong vertical displace- ment along the block structures (uplift/subsidence) a composite hilly relief of this area was created. The Rakovica sands are composed of a single uniform de- posit, with a maximum thick ness of 3 m at the studied exposure (Fig. 1). The section has no visible layers. It is dominated by gray- yellow quartz sands with unevenly distributed fossils, and poorly- cement ed sands with hard sandstone lenses. The faunal assem- blage of this locality is rich, both in terms of numbers and diver sity of taxa, consisting of gastropods, bivalves, foraminifers, corals, bryozoans, ostracods, annelids, crabs (fragments of pincers), sca- phopods, sea urchins (spine fragments), fish otoliths and shark teeth. Biostratigraphically, based on foraminifers, the sands from the Rakovica stream belong to the Elphidium crispum–Ammonia viennensis Ecozone, the upper Badenian Stage (PETROVIĆ, 1985; PETROVIĆ & ŠUMAR, 1990). The main taxa among fo- rams includes: Elphidium cris pum, El. fichtelianum, Ammonia vien nensis, Borelis melo, Spiroli na austriaca, Asterigerinata plan orbis, Cibicidoides cf. lobatu lus, Lenticulina sp., Quinquelo­ culina sp., Planostegina sp. More over, numerous ostracods in- cluding Aurila punctata, A. cicatricosa, Cle to cythereis haidin­ geri, Falunia plicatula, Cnestocythere trun cata, Costa edwardsi, Tenedocythere sulcatopunctata, Pokor nyella deformis, Bairdo­ pipilata subdeltoidea, Loxoconcha has tata, Callistocythere ca­ nalicullata, etc. were found. A similar exposure is located in the Belgrade City centre, beneath the Belgrade Fortress, where the sands lie next to the Miocene marine reef „Kalemegdan” (a geo- heritage site), mostly composed of coralline alga Lithothamnium (EREMIJA, 1977). Although it forms a part of the same geologi- cal unit (the Rakovica sands), the faunal assemblage from the Bel- grade Fortress is not nearly as rich as in the Rakovica stream. Nevertheless, it would be interesting to sample this exposure too, but since it is located in the zone of the protected geoheritage site, this was impossible. Since the exposure is relatively small and lithologically uni- form, and because it will probably become a geoheritage site soon, we did not want to damage the site by taking multiple sam- Figure 1. A simplified geographic position and location of the studied area. Figure 2. A general stratigraphic column of the Miocene of Belgrade area (modified after SCHWARZHANS et al., 2015). A gray rectangle marks the stratigraphic position of the studied sediments. Age of the Badenian regional stage based on HOHENEGGER et al., 2014. G eologia C roatica Ganić et al.: Traces of drilling predation in the Upper Badenian (Middle Miocene) molluscs from the Rakovica stream (Belgrade) 207 ples. Therefore, after we conducted a detailed examination of the entire exposure, we decided to collect only a single sample from the small surface of the vertical profile, located between 0.5 and 1 m from the accessible foothill (Fig. 2). 3. MATERIALS AND METHODS The sample (bulk sediment) of about 10 kg was brought back to the laboratory for study. The sediment was processed using fine­ mesh screens (0.15 to 0.80 mm) with seven sieves. Bivalves and gastropods were retained on screens 0.8; 0.7; and 0.6 mm. Traces of drilling have been identified on shells larger than 0.7 mm. All mollusc specimens were counted and examined for evidence of drilling predation under a binocular magnifier (REICHERT), with 20 x magnification for small and 2 x for large specimens. Initially, we counted the abundance at the different taxonomic levels (species, family, class and total). The frequencies of preda- tor-prey interactions in the sample were analysed by calculating drilling frequency (DF), measuring the rate of prey mortality due to drilling predation, by dividing the number of specimens that contain at least one successful predation trace (DS) by the total number of specimens in the sample (n). For bivalves, the total number of specimens had to be corrected by dividing it by 2, in order to understand disarticulation of the shells after death. DF was calculated at family level (Lower Taxon Frequency accord- ing to KOWALEWSKI, 2002), but also at class and total assem- blage levels (Assemblage Frequency according to KOWA LEW- SKI, 2002). Escalation parameters, (estimates that provide some measure of the predator’s failure, or rather a prey’s ability to re- sist predators), were assessed by calculating a relative frequency of failed predator attacks. This frequency is often refer red to as prey effectiveness (PE), and is calculated by dividing the number of traces of unsuccessful attacks (ID) (incomplete, repaired and healed drill holes) by the total number of drilling attempts (ID plus number of complete drill holes, D) (KOWA LEWSKI, 2002; SAWYER & ZUSCHIN, 2011). Chi-squared tests were used to compare drilling frequencies (on total-assemblage, class and fami ly levels) and prey effectiveness (on total-assemblage and class levels) from the Rakovica sample with those published for Paratethys and other Miocene basins (HOFFMEISTER & KOWA LEW SKI, 2001; KELLEY & HANSEN, 2006; SAWYER & ZUSCHIN, 2011). At family level, only data for the Central Paratethys (SAWYER & ZUSCHIN, 2011) were available. A 5% significance criterion (α = 0.05) was applied to our statistical analysis, and an online interactive calculation tool was used for the chi-square tests (PREACHER, 2001). 4. RESULTS A total of 2301 mollusc shells, consisting of 1617 gastropods and 684 bivalves, were recovered from the sample. 56 gastropod and 14 bivalve species (Tables 1 and 2) were identified. Evidently, gastropods dominated the sample both in terms of species rich- ness and biomass. However, bivalves were more frequently at- Table 1. The list of species recovered from Rakovica sample, with numbers of specimens/valves (n) in parantheses. Class Species (n) Gastropoda Acirsa drevermani (18), Acteocina lajonkaireana (15), Alaba costella anomala (127), Alvania oceani (31), Alvania productilis (1), Ancilla glandiphormis (25), Ath- leta ficulina (67), Athleta pirulaeformis (5), Athleta rarispina (10), Bittium spina (95), Buccinum limatum (11), Calyptrea chinensis (7), Cerithiopsis bilineata (1), Cerithiopsis opaca (15), Cerithium crenathum (17), Cerithium procrenatum (6), Clavatula capgrandi (10), Clavatula asperulata (29), Conus dujardini (21), Conus fiscocingulathus (96), Conus granularis (2), Euthria intermedia (5), Fusinus subrugosa (6), Fusinus valenciensis (1), Hydrobia hoernesi (116), Mangelia perpulchra (25), Marginella cratoformis (27), Marginella haueri (7), Marginella minuta (10), Mitrella bitneri (32), Mitrella fillisae (6), Mitrella hilberi (53), Murex granularis (3), Nassa badensis (81), Nassarius seraticosta (78), Natica millepunctata (22), Natica redempta (27), Neritina picta (21), Odontostoma plicatum (10), Oxistella patu- la orientalis (96), Piramidella plicosa (12), Retusa truncatula (69), Ringicula bucinae (6), Rissoa turricula (42), Rissoina pusilla (8), Sandbergeria perpusila (125), Scala tenuicosta (2), Scaphander lignaris (1), Strioterebrum basteroti (7), Terebra fuscata (2), Turicula miocenica (13),Turris disjuncta (18), Turritella badensis (6), Turritella bicarinata (2), Turritella subangulata (41), Turritella subangulata pulchra (8) Bivalvia Glycymeris glycymeris (91), Cardites partschi (35), Euripicardium multicostatum (22), Clausinella basteroti (22), Loripes dentatus (158), Anadara diluvii (62), Nu- cula (Nucula) nucleus (17), Gouldia minima (168), Tellina donacina (41), Linga (Linga) collumbela (14), Corbula carinata (24), Parvicardium papillosum (2),Cubi- tostrea digitalina (9), Glycymeris obtusata (19) Table 2. Taxonomic summary of drill hole data from mollusks from the Rakovica site for the overall assemblage, classes and families. n = abundance (corrected to account for disarticulated bivalves), DS = number of shells with complete drill holes, D = number of complete drill holes, ID = number of incomplete drill holes, DF = drilling frequency, PE = prey effectiveness. n DS D ID DF (%) PE (%) Total assemblage 1959 214 226 19 10.92 7.76 Class Gastropoda 1617 156 168 13 9.65 7.18 Class Bivalvia 342 58 58 6 16.96 9.38 Gastropod families Epitoniidae 18 1 1 0 5.55 0 Cylichnidae 16 3 3 0 18.75 0 Litiopidae 127 13 13 0 10.24 0 Rissoidae 84 11 11 2 13.09 15.38 Olividae 25 5 5 0 20 0 Volutidae 82 0 0 0 0 – Cerithiidae 118 17 17 2 14.41 10.53 Buccinidae 16 2 2 0 12.5 0 Calyptraeidea 7 2 2 1 28.57 33.33 Cerithiopsidae 16 3 3 0 18.75 0 Turridae 82 6 7 0 7.32 0 Conidae 119 4 6 0 3.36 0 Fasciolariidae 7 1 1 0 14.28 0 Hydrobiidae 116 7 7 0 6.03 0 Columbellidae 38 4 4 0 10.53 0 Mitridae 53 7 7 0 13.21 0 Muricidae 3 0 0 0 0 – Nassaridae 159 8 8 0 5.03 0 Naticidae 49 3 3 1 6.12 25 Neritidae 21 0 0 0 0 – Pyramidellidae 22 5 5 2 22.73 28.57 Trochidae 96 2 2 1 2.08 33.33 Retusidae 69 22 31 0 31.88 0 Ringiculidae 6 0 0 0 0 – Scaliolidae 127 16 16 2 12.6 11.11 Terebridae 9 0 0 0 0 – Turritellidae 75 13 13 1 17.33 7.14 Marginellidae 44 1 1 1 2.27 50 Clavatulidae 13 0 0 0 0 – Bivalve families Glycymerididae 55 3 3 1 5.45 25 Carditidae 17.5 4 4 0 22.86 0 Cardiidae 12 1 1 0 8.33 0 Veneridae 95 11 11 1 11.58 8.33 Lucinidae 86 31 31 2 36.05 6.06 Arcidae 31 5 5 0 16.13 0 Nuculidae 8.5 2 2 0 23.53 0 Tellinidae 20.5 0 0 2 0 100 Corbulidae 12 0 0 0 0 – Ostreidae 4.5 1 1 0 22.22 0 G eo lo gi a C ro at ic a Geologia Croatica 69/2208 tacked than gastropods, since drilling frequency was higher in the former (16.96%) than the latter (9.65%). In terms of prey ef- fectiveness, bivalves (9.38%) were able to resist predator attacks more successfuly than gastropods (7.18%). The most abundant gastropod families (Table 2), with more than 115 specimens in the sample, were Nassaridae (159), Liti­ opidae (127), Scaliolidae (127), Conidae (119), Cerithiidae (118) and Hydrobiidae (116). With numbers between 96 and 44, the fol- lowing families were also relatively numerous: Trochidae (96), Rissoidae (84), Volutidae (82), Turridae (82), Turritellidae (75), Mitridae(53), Naticidae (49) and Marginellidae (44). With abun- dances between 38 and 13, Columbellidae (38), Olividae (25), Py ramidellidae (22), Neritidae (21), Epitoniidae (18), Cylichnidae (16), Buccinidae (16), Cerithiopsidae (16), Cylichnidae (16) and Clavatulidae (13), were much less abundant. Six families were represented by less than 10 shells – Terebridae (9), Calyptraeidea (7), Fasciolariidae (7), Ringiculidae (6) and Muricidae (3). Six of 29 gastropod families were never drilled, but only two of those had abundances greater than 20 (Volutidae and Neritidae). Among the attacked families, the DF ranged from 2.08% (Tro­ chidae) to 31.88% (Retusidae). Incomplete drill holes were ob- served in nine families, with PE ranging from 7.14% (Turritelli­ dae) to 50% (Marginellidae); however, there were only one or two incomplete drillings per family. Most families were attacked be- tween one and 19 times (succesful plus unsuccesful attacks); the notable exception were Retusidae, with 31 complete, but no in- complete drill holes. Multiple drill holes were found in three fami- lies (Turridae, Conidae, Retusidae). Three bivalve families dominated the sample (Table 2) – Ve­ ne ridae (95), Lucinidae (86) and Glycymerididae (55). The Arci­ dae were represented by 31 specimens, Tellinidae with 20.5 and Corbulidae had 12; Nuculidae and Ostreidae had less than 10 spe cimens (8.5 and 4.5, respectively). The abundance data was corrected to account for disarticulated bivalves. Only one of 10 bivalve families was never drilled (Corbulidae), while one family showed traces of only two failed drilling attempts (Tellinidae). Among those drilled, the DF ranged from 5.45% (Glycymeridi­ dae) to 36.05% (Lucinidae). Incomplete drill holes were pre sent in four families, with the PE ranging from 6.06% (Lucinidae) to 100% (Tellinidae); however, as in gastropods, there were no more than two IDs per family. Most families were attacked (succesful plus unsuccesful attacks) five times or less, except for the Luci­ nidae and Veneridae, with 33 and 13 attacks, respectively. No multiple drill holes were observed on bivalve shells. Based on our analysis (Table 3) the total (mollusc) assem- blage-level drilling frequency for the Rakovica sample (10.92%) is lower than the DF reported for the Paratethys province (15.2%) and significantly (two/three­fold) lower than those reported for the Boreal (22.4%), southeastern North Atlantic (23.5%) and US Atlantic coast (34.4%). However, the total assemblage-level DF for Rakovica is higher than the value reported for Central Para- tethys (7.5%). These differencies were all statistically significant, as showed by the chi-square tests. Comparisons of the assem- blage-level DF for gastropods quite gave similar results and all differences were also statistically significant (Table 3). Alterna- tively, the assemblage-level DF values for bivalves, showed that only two comparisons were statistically significant. The DF for the Rakovica sample (16.96%) is almost two times higher than the DF reported for the Lower/Middle Miocene of the Central Paratethys (8.6%). Conversly, the DF for bivalves is two-fold lo- wer than the DF reported for the US Atlantic coast (34.1%). Ex- amination of the drilling frequencies for the Rakovica and Central Paratethys samples at the family­level, statistically significant dif­ ferences only existed for three taxa (Table 4). Among the gastro- pods, the Cerithiidae were drilled almost twice as frequently in the Rakovica sample than in Central Paratethys (DFs are 14.41% and 7.6%, respectively). Among bivalves, the Veneridae show similar differences (DF in Rakovica sample is 11.58%, com pared to 7.6% in Central Paratethys). Lucinidae were drilled five times more frequently in the Rakovica sample than in the Central Pa- ratethys province (DFs were 36.05% and 7.1%, respectively). Table 3. Statistical comparisons of total and phylum assemblage-level drill frequencies using chi-squared test. c2 = chi-square. Bold p-values are statistically signifi- cant at α = 0.05. Level Sample Drilled Undrilled DF (%) c2 p-value Source Total assemblage Rakovica 214 1745 10.92 – – Present study Central Paratethys 2324 28636 7.5 30.24 < 0.00001 SAWYER & ZUSCHIN, 2011 Boreal 386 1336 22.4 88.71 < 0.00001 HOFFMEISTER & KOWALEWSKI, 2001 Paratethys 156 868 15.2 11.5 0.000696 HOFFMEISTER & KOWALEWSKI, 2001 Southeastern North Atlantic 27 88 23.5 16.67 0.000045 HOFFMEISTER & KOWALEWSKI, 2001 US Atlantic Coast 11483 21945 34.4 458.99 < 0.00001 KELLEY & HANSEN, 2006 Gastropoda Rakovica 156 1461 9.65 – – Present study Central Paratethys 1596 20698 7.2 13.75 0.000209 SAWYER & ZUSCHIN, 2011 Boreal 284 875 24.5 111.72 < 0.00001 HOFFMEISTER & KOWALEWSKI, 2001 Paratethys 84 515 14 8.66 0.003242 HOFFMEISTER & KOWALEWSKI, 2001 Southeastern North Atlantic 22 45 32.8 36.59 < 0.00001 HOFFMEISTER & KOWALEWSKI, 2001 US Atlantic Coast 4048 7573 34.8 415.48 < 0.00001 KELLEY & HANSEN, 2006 Bivalvia Rakovica 58 284 16.96 – – Present study Central Paratethys 726 7747 8.6 28.562 < 0.00001 SAWYER & ZUSCHIN, 2011 Boreal 102 461 18.1 0.196 0.657969 HOFFMEISTER & KOWALEWSKI, 2001 Paratethys 72 353 16.9 0 1 HOFFMEISTER & KOWALEWSKI, 2001 Southeastern North Atlantic 5 43 10.4 1.33 0.248805 HOFFMEISTER & KOWALEWSKI, 2001 US Atlantic Coast 7435 14372 34.1 44.167 < 0.00001 KELLEY & HANSEN, 2006 Ganić et al.: Traces of drilling predation in the Upper Badenian (Middle Miocene) molluscs from the Rakovica stream (Belgrade) 209 G eologia C roatica Due to the small number of unsuccessful drilling attempts (ID) in our sample and the lack of the appropriate comparative data, prey effectiveness could only be studied and tested for sta- tistical significance in a very limited manner (Table 5). On the total assemblage-level, molluscs from Rakovica were less able to resist predator attacks than those from the Boreal and Parathethys province in general. When compared to Central Paratethys, the frequency of failed predator attacks on gastropods alone show ed only a slightly lower value (PEs for Central Paratethys and Ra- kovica were 7.4% and 7.18%, respectively). Compared to Central Paratethys (6.3%), bivalves from Rakovica showed a high er PE value (9.38%). 5. DISCUSSION The study of predator-prey interactions by calculating and com- paring drilling frequency and prey effectiveness data is a com- mon and widely used approach in fossil and modern environ- ments. Although the drilling data for lower taxa (species, genus, family) probably provide the most biologically meaningful ana- lysis of predator-prey interaction in the fossil record (KOWA LEW- SKI, 2002), assemblage-level analysis is more useful in the pragma- tic sense. This is because assemblage-le vel data can be computed for any fossil sample and thus can be used for analytical compar- isons throughout the fossil record. However, assemblage-level analyses are quite problematic, since they often mix habitats, combine prey with different morphological and ecological chara- cteristics and do not take into account other, fine­scale palaeo- ecological data. A detailed review of the advantages and dis- advantages of this method is outwith the scope of this paper, but can be seen elsewhere (KOWALEWSKI, 2002; LEIGHTON, 2002; VERMEIJ, 2002). In this paper, the assemblage-level, class-level and family- level drilling data was calculated for the Upper Badenian strata at the Rakovica stream in Belgrade, and compared to available data for Miocene provinces (Central Paratethys, Paratethys in general, Boreal, Southeastern North Atlantic and United States Atlantic Coast). This study is the first of this kind to be conduct ed on a sample originating from the Southern part of the Central Pa- ratethys province. As some of the published studies have already shown (HAN- SEN & KELLEY, 1995; HOFFMEISTER & KOWALEWSKI, 2001; SAWYER & ZUSCHIN, 2011), predation trace data from a single horizon, locality or province can vary significantly, due to the problem of spatial variation. This problem can possibly be solv ed by calculating average values derived from multiple sam- ples (HOFFMEISTER & KOWALEWSKI, 2001) and was used in this study. For the reasons briefly explained above, it was not pos- sible to take multiple and spatially separated samples from the Ra- kovica sands, therefore our sample may misrepresent the locality. However, the sample is relatively large (almost 2000 spe cimens) and there is a good chance that the sampling error is not high. Based on the studies of lithological composition, presence of the diverse marine fauna (e.g. molluscs, foraminifers, and ostra- cods), and local palaeogeographical setting, it can be con clud ed with reasonable certainty that the Rakovica sands represent ed a shallow marine (littoral) environment during Late Badenian time (MARO VIĆ et al., 2007). The waters in this part of Central Pa- ratethys were warm, since they were positioned in the subtropical climatic zone (KOVAČ et al., 2007). The presence of the warm- loving benthic fish (family Gobiidae) at the Rakovica stream (NELSON, 2006) supports this view. In this shallow water envi- ronment, the salinity was normal, with an occasional brackish in- fluence, confirmed by discovery of the mixed assemblage of typi­ cal marine (corals, sea urchins) and euryhaline fau na (Cerithidae, Hydrobiidae and Lucinidae). Discovery of scarce fragments of corals and echinoderms also indicates more dynamic conditions, water fluctuations and higher energy of water than the salinity. Some ostracods (Callistocythere, Aurila, Loxoconcha) sug gest the upper part of the infralittoral zone (up to 40m, ZORN, 2003). The Rakovica ostracod fauna lived in shallow (about 50 m deep), warm, and clear waters, connected to a deeper sea, occasionally exposed to freshwater inflows. The finding of the foraminifer ge- nus Borelis indicates very warm water and implies tropical to subtropical water temperatures. A similar indication was given by SZCZECHURA & ABD-ELSHAFY (1988) and RUNDIĆ (1992) from the Badenian deposits of North Africa and NE Bos- nia and Herzegovina. Somewhere, a very steep coast could be responsible for the extreme mixing of the marine fauna and also explains some brackish/terrestrial faunal elements at the sam- pling locality. Table 4. Statistical comparisons of family-level drill frequencies using chi- squared test. Bold p-values are statistically significant at α = 0.05. Comparative data for the Central Paratethys comes from SAWYER & ZUSCHIN (2011). Family Rakovica DF (%) Central Para- tethys DF (%) c2 p-value Rissoidae 13.09 10.9 0.398 0.53 Cerithiidae 14.41 7.6 6.689 0.0097 Turridae 7.32 8.6 0.119 0.730122 Nassaridae 5.03 6.8 0.742 0.39 Pyramidellidae 22.73 9.4 2.377 0.12 Naticidae 6.12 12.9 1.88 0.17 Veneridae 11.58 6 4.065 0.043781 Lucinidae 36.05 7.1 58.459 < 0.00001 Arcidae 16.13 4.7 3.776 0.051993 Table 5. Statistical comparisons of total assemblage and phylum-level prey effectiveness using chi-squared test. Bold p-values are statistically significant at α = 0.05. Level Sample D ID PE (%) c2 p-value Source Total assemblage Rakovica 226 19 7.76 – – Present study Central Paratethys 2324 165 6.6 0.45 0.502335 SAWYER & ZUSCHIN, 2011 Boreal 386 77 16.63 10.768 0.001033 HOFFMEISTER & KOWALEWSKI, 2001 Paratethys 156 3 18.87 6.449 0.011101 HOFFMEISTER & KOWALEWSKI, 2001 Gastropoda Rakovica 168 13 7.18 – – Present study Central Paratethys 1596 58 7.4 5.926 0.014919 SAWYER & ZUSCHIN, 2011 Bivalvia Rakovica 58 16 9.38 – – Present study Central Paratethys 726 107 6.3 4.466 0.034576 SAWYER & ZUSCHIN, 2011 G eo lo gi a C ro at ic a Geologia Croatica 69/2210 Most of the identified bivalve taxa lived as infauna, buried in soft substrate on the seabed. They were sustained by filtration of suspended food, such as plankton, algae and other water-borne nutrients and particles (Cardiidae, Veneridae, etc.), or by chemo- synthetic bacteria (in the case of Lucinidae). Carnivorous gastro- pods dominated the sample, comprising more than fifty percent (52.32%) of all gastropod specimens, and also a large por tion of the total sample (43.19%). The most numerous of these were Nas­ saridae, Conidae, Turridae and Volutidae. Two families of preda- tors (Naticidae and Muricidae) that were most likely responsible for the drill holes were observed in our sample. These gas tropods gain access to the soft parts of the prey by the radula. Our results on drilling intensity showed that bivalves from the Rakovica sands were attacked almost twice as frequently as gastropods (Table 2). This is probably related to the specific mode of life of these molluscs (infauna) and their enemies, naticids. Al- ternatively, prey effectiveness data (Table 2) indicates that the bivalves were more successful at resisting predator attacks. Again, this may be because most bivalves from the sample lived as infauna, which is hard to reach for predators (BARNES et al., 1988). In contrast, most of the identified gastropod prey lived as epifauna; and this makes them less accessible to infaunal preda- tory naticids, which constitute the majority of drilling predators in our study. Most of the mollusc shells showed only traces of a single (suc- cessful) drilling attempt. Multiple drill holes were found only in three gastropod families (Turridae, Conidae and Retusidae) and only on a single specimen from each group. Several scenarios could explain these cases of multiple drilling. We could imagine more than one predator attacking the prey at the same time, but it could also be the case that individual drillings were made at dif- ferent times (KITSCHELL, 1986), by the same or a different preda- tor. Also, a predator could have been attacking an empty shell (HOFFMAN et al., 1974). Unfortunately, experimen tal studies (HUTCHINGS & HERBERT, 2013) failed to solve this problem. Six gastropod families (Volutidae, Muricidae, Neritidae, Ring i culidae, Terebridae, Clavattulidae) and one bivalve family (Corbulidae) were never drilled; also, Tellinidae showed only failed drilling attempts. This is probably because many of these lived buried deep in the sediment similar to the representatives of families Tellinidae, Corbulidae and Volutidae and relatively thick shells (Volutidae). The most numerous family (82 speci- mens) with no drillings on their shells were the Volutidae; inter- estingly, one specimen that belonged to this family as reported by SAWYER & ZUSCHIN (2011) also had no drill holes (although this could be just a coincidence, considering the small sample). Here, naticids and muricids were considerably less abundant, forming 2.65% percent of the total sample. With just three spec- imens, the muricids were particularly rare. However, in the Cen- tral Paratethys sample published by SAWYER & ZUSCHIN (2011), these two predator families together comprise only 1.6% of the total mollusc sample. A considerably higher percentage (4.2%) of these predators was observed by KELLEY & HANSEN (2006) in the U.S. Atlantic Costal Plains. Of course, we should keep in mind that these comparative figures were made on the basis of multiple samples from localities with different palaeoeco- logical and stratigraphic settings. Moreover, 49 naticid spe cimens were attacked; all of these drillings were probably made by the members of the same taxa, because naticids are known for can- nibalism (CHATTOPADHYAY et al., 2014a).Traces of predation muricids (a specific shape of drill hole; CARRIKER, 1981) were only determined on four gastropod shells. For a more pre cise as- sessment, a more detailed analysis should be made (ZŁO TNIK, 2001, CHATTOPADHYAY et al., 2014a), but it seems highly un- likely that the lack of infaunal prey was the cause of cannibalism among the naticids. Comparisons of total assemblage-level and class-level dril- ling frequencies (Table 3) showed that the value for the Rakovica stream was significantly lower than the data reported for the Bo- real, Paratethys in general, Southeastern North Atlantic and US Atlantic Coast. However, the value reported for the Central Pa- ratethys was lower than the value reported here. These differ- ences could be due to the fact that samples come from different stratigraphic settings: the Rakovica sands is Upper Badenian, the Central Paratethys sample comes from the Karpatian – Badenian, Paratethys sample is from Burdigalian – Serravallian, and Boreal, Southeastern North Atlantic and US Atlantic Coast are from Mid- dle – Late Miocene. Other abiotic factors could also be responsi- ble – differences in depth, seabed composition, salinity, tempera- ture etc. Alternatively, the Central Paratethys represent ed an epicontinental sea with specific evolution, which lasted for less time than the evolutions of other basins used in our comparative analysis (RÖGL, 1998); therefore, we can assume that fossil com- munities were also distinctive (HARZHAUSER et al., 2003). This also means that communities of the Central Paratethys (SAWYER & ZUSCHIN, 2011) and Rakovica stream (which rep- resents its southeastern part of Central Paratethys) are more simi- lar to each other than to other basins; indeed, differences in DF values seem to confirm this. Also, we think that the higher DF in this study, relative to the Central Paratethys, is in accordance with the littoral character of the locality, since there is a negative cor- relation between drilling frequency values and increasing depth (WALKER, 2001; TOMAŠOVYCH & ZUSCHIN, 2009). It is believed that drillings are more frequent in warm, tropical envi- ronments, where mollusc numbers and diversity are far greater than in cold waters (VERMEIJ & ROOPNARINE, 2001, CHAT- TOPADHYAY et al., 2014b). Since there is actually a higher DF in the relatively cold realms of the Atlantic and Boreal, tempera- ture was not the main cause for the observed differences. In the case of the Central Paratethys (SAWYER & ZUSCHIN, 2011) the low abundance of drilling predators may be a very good reason for low DF. Primarily, we must keep in mind that we are compar- ing a single locality with the cumulative data for entire marine realms, which included many datasets, from different types of sediments that were formed in different abiotic environments. This fact alone could be responsible for the large proportion of the differences observed. Family-level DF values in this study were compared with data from the Central Paratethys (Table 4) and showed that Ce­ rithidae, Veneridae and Lucinidae were drilled significantly more frequently in our sample; all of these taxa represent non-predatory (grazers, suspension feeders and chemosymbionts, respectively) and euryhaline taxa. Their relatively high abundances and high DFs (especially for Veneridae and Lucinidae) could mean that they were preferred by predators; these taxa would have been more available in shallow waters with lowered salinity along the south- eastern rim, than in the Central Paratethys in general. Despite the methodological circumstances mentioned earlier, our comparisons of escalation data (Table 5) showed that mol- luscs from the Rakovica sands were possibly less successful in resisting predator attacks than those from Boreal and Paratethys provinces. This could be caused by biotic factors (such as a rela- tively higher number of predators etc.) or abiotic factors (such as a relative depth, seafloor composition etc.). When class­level es- G eologia C roatica Ganić et al.: Traces of drilling predation in the Upper Badenian (Middle Miocene) molluscs from the Rakovica stream (Belgrade) 211 calation data from all parts of the Central Paratethys were com- pared with the SE parts in Belgrade, almost identical figures for gastropods were observed, while apparently bivalves from this study were more successful in resisting predator attacks. This too could be explained by the number of biotic and abiotic factors, or simply by a sampling artefact. 6. CONCLUSION The analysis of predator-prey interactions in the southeastern part of Central Paratethys during the Badenian was conducted on the sample from the Rakovica stream locality (the Rakovica sands) in Belgrade. The fossiliferous sands were deposited in the littoral zone of a warm sea. The salinity in this part of the basin was nor- mal, with an occasional brackish influence, as indicated by the mixing of typical marine and euryhaline faunas. From the bio- stratigraphic point of view, the Rakovica sands belong to the El­ phidium crispum – Ammonia viennensis Ecozone (Upper Bade- nian). Statistical analysis of the rich gastropod and bivalve communities showed that bivalves were more frequently attacked by drilling predators than gastropods. Bivalves were also more successful in resisting predator attacks, probably due to the biotic factors, such as the dominant infaunal mode of life. Predatory gastropods such as the Naticidae and Muricidae families were mostly responsible for the drill holes examined in our study, but muricids were particularly rare. In this study, comparisons with other Miocene provinces showed that the details of predator-prey interactions, matched (with minor deviations) the Central Paratethys, which indicates relative uniform palaeoecological conditions in the basin. How- ever, there are relatively small differences between our results and data from all the afore-mentioned parts of the Central Paratethys province. At the Rakovica sands, there are higher DF and PE val- ues because we were dealing with a single sample (environment), with specific local conditions such as temperature, salinity, depth etc. In contrast, all the other data used for comparison were cu- mulative, based on samples from different en vironments – littoral versus sublittoral, and estuaries versus fully marine settings and different strata. Other provinces mentioned here, show different results, both in terms of drilling intensities and escalation data. Finally, we concluded that there is no general pattern for the interpretation of the palaeoecological conditions which have been responsible for the formation of different populations of fossil organisms and their predator – prey relationships. ACKNOWLEDGMENT M. GANIĆ, LJ. RUNDIĆ and S. KNEŽEVIĆ wish to thank the Ministry of Education and Science, Republic of Serbia (Project no. 176015) for supporting their contribution to this study. We also wish to express special gratitude to the reviewers for their constructive comments and suggestions that greatly improved an earlier version of the manuscript. REFERENCES BARNES, R.S.K., CALLOW, P. & OLIVE, P.J.W. (1988): The Invertebrates: A New Synthesis.– Blackwell Scientific Publications, 118–133. BAUMILLER, T.K. (1996): Boreholes in the Middle Devonian blasoid Heteroschisma and their implications for gastropod drillig.– Palaeogeography, Palaeoclimatology, Palaeoecology, 123, 343–351. CARRIKER, M.R. (1981): Shell penetration and feeding by Naticacean and Muricacean predatory Gastropods: A Synthesis.– Malacologia, 20/2, 403–422. CHATTOPADHYAY, D. & DUTTA, S. (2013): Prey selection by drilling predators: A case study from Miocene of Kutch, India.– Palaeogeography, Palaeoclimatology, Palaeoecology, 374, 187–196. doi: 10.1016/j.palaeo.2013.01.016 CHATTOPADHYAY, D., SARKAR, D., DUTTA, S. & PRASANJIT, S.R. (2014a): What controls cannibalism in drilling gastropods? A case study on Natica tigrina.– Pal- aeogeography, Palaeoclimatology, Palaeoecology, 410, 126–133, doi: 10.1016/j. palaeo.2014.05.037 CHATTOPADHYAY, D., ZUSCHIN, M. & TOMAŠOVYCH, A. (2014b): Effects of high risk environment on edge-drilling behavior: Inference from Recent bivalves from the Red Sea.– Paleobiology, 40, 34–49, doi: 10.1666/13024 EREMIJA, M. (1977): Miocen u oblasti ekshumiranog mezozojsko-serpentinskog pale- oreljefa, od Rakovice do Ripnja. [Miocene in the exhumed Mesozoic­serpentine area of palaeorelief, from Rakovica to Ripanj – in Serbian].– In: PETKOVIĆ, K. (eds.): Géologie de le Serbie. Stratigraphie II/3: 145–154. EREMIJA, M. (1987): Badenski kat [Badenian – i Serbian].– In: ANĐELKOVIĆ, M. (eds.): La Géologie des Environs de Belgrade I, Géologie et Géodinamique, 151–172. HANSEN, T.A. & KELLEY, P.H. (1995): Spatial Variation of Naticid Gastropod Preda- tion in the Eocene of North America.– Palaios, 10, 268–278. doi: 10.2307/3515257 HARZHAUSER, M., MANDIC, O. & ZUSCHIN, M. (2003): Changes in Paratethyan marine mollusks at the Early/Middle Miocene transition: Diversity, palaeogeogra- phy and palaeoclimate.– Acta Geologica Polonica, 53, 323–339. HOFFMAN, A., PISERA, A. & RYSZKIEWICZ, M. (1974): Predation by muricid and naticid gastropods on the Lower Tortonian mollusks from the Korytnica Clays.– Acta Geologica Polonica, 24, 249–260. HOFFMEISTER, A.P. & KOWALEWSKI, M. (2001): Spatial and Environmental Varia- tion in the Fossil Record of Drilling Predation: A Case Study from the Miocene of Central Europe.– Palaios, 16, 566–579. doi: 10.1669/0883-1351(2001)016%3C0566: SAEVIT%3E2.0.CO;2 HOHENEGGER, J., ĆORIĆ, S. & WAGREICH, M. (2014): Timing of the Middle Mio cene Badenian Stage of the Central Paratethys.– Geologica Carpathica, 65/1, 55–66, doi: 10.2478/geoca-2014-0004 HUTCHINGS, J.A. & HERBERT, G.S. (2013): No honor among snails: Conspecific competition leads to incomplete drill holes by a naticid gastropod.– Palaeogeog- raphy, Palaeoclimatology, Palaeoecology, 379–380, 32–38, doi: 10.1016/j.palaeo. 2013.04.003 IVKOVIĆ, A., VUKOVIĆ, A., NIKOLIC, J., KOVAČEVIĆ, D., PALAVESTRIĆ, LJ., PETROVIĆ, LJ., JOVANOVIĆ, O., TRIFUNOVIĆ, I. & SEBINOVIĆ, LJ. (1966): Osnovna geološka karta SFRJ 1:100000, List Pančevo L34­114 [Basic Geological Map of SFRY 1:100000, Pančevo sheet – in Serbian].– Geološki zavod, Beograd, Savezni geološki zavod, Beograd. JOVANOVIĆ, G. (2002): Paleoecological and taphonomical features of Badenian fauna of Sladinci near Golubac.– Proceedings of Geoinstitute, 37, 135–150. KELLEY, P.H. & HANSEN, T.A. (1993): Evolution of the Naticid Gastropod Predator- Prey System: An Evaluation of the Hypothesis of Escalation.– Palaios, 8, 358–375. doi: 10.2307/3515266 KELLEY, P.H. & HANSEN, T.A. (1996): Recovery or the naticid gastropod predator- prey system from the Cretaceous-Tertiary and Eocene-Oligocene extinctions.– Geo- logical Society Special Publication, 102, 373–386. doi: 10.1144/GSL.SP.1996.001. 01.27 KELLEY, P.H. & HANSEN, T.A. (2006): Comparisons of class- and lower taxon-level patterns in naticid gastropod predation, Cretaceous to Pleistocene of the U.S. Coastal Plain.– Palaeogeography, Palaeoclimatology, Palaeoecology, 236, 302– 320. doi: 10.1016/j.palaeo.2005.11.012 KITCHELL, J.A. (1986): The evolution of predator-prey behavior: Naticid gastropods and their molluscan prey.– In: NITECKI, M. & KITCHELL, J.A. (eds.): Evolution of Animal Behavior: Paleontological and Field Approaches Oxford Univerity Press, Oxford, 88–110. KNEŽEVIĆ, S. (1989): Prinove iz ponta belopotočkog rova [New specimen from the Pontian of Beli Potok trough – in Serbian, Summary in English].– Ann. Geol. Pen- nins. Balk., 53, 373–382. KOHN, A.J. (1983): Feeding Biology of Gastropods.– In: SALEUDDIN, A.S.M. & WIL- BUR, K.M. (eds.): The Mollusca, edit. 5, Physiology 2, 1–63. doi: 10.1016/B978- 0-12-751405-5.50009-9 KOVÁČ, M., ANDREYEVA­GRIGOROVICH, A., BARJAKTAREVIĆ, Z., BRZOBO- HATY, R., FILIPESCU, S., FODOR, L., HARZHAUSER, M., NAGZMAROSY, A., OSZCZYPKO, N., PAVELIĆ, D., RÖGL, F., SAFTIĆ, B., SILVA, L. & STU- DENCKA, B. (2007): Badenian evolution of the Central Paratethys Sea: Paleoge- ography, Climate and Eustatic Sea-level Changes.– Geologica Carpathica, 58, 6, 579–606. KOWALEWSKI, M. (2002): The Fossil Record of Predation: An Overview of Analyti- cal Methods.– In: KOWALEWSKI, M. & KELLEY, P.H. (eds.): The fossil Record of predation. Paleontological Special Papers, 8, Yale University, New Haven, 3–42. LEIGHTON, L.R. (2002): Inferring predation intensity from the marine fossil record.– Paleobiology, 28, 328–342. LUKOVIĆ M. (1922): Facije drugog mediteranskog kata u okolini Beograda [Second Mediterranean facies from Belgrade surroundings – in Serbian].– Ann. Geol. Pen- nins. Balk., 7/1, 22–41. G eo lo gi a C ro at ic a Geologia Croatica 69/2212 MAROVIĆ, M., TOLJIĆ, M., RUNDIĆ, LJ. & MILIVOJEVIĆ, J. (2007): Neoalpine Tectonics of Serbia.– Serbian Geological Society, Ser. Monographie, 87 p., ISBN 978-86-86053-03-9 NELSON, J.S. (2006): Fishes of the World.– Department of Biological Sciences, Uni- versity of Alberta, Canada, 601 p. doi: 10.1002/9781119174844 PAVLOVIĆ, P.S. (1890): Mediteranska fauna u Rakovici [Mediteranian fauna from the Rakovica – in Serbian].– Ann. Geol. Pennins. Balk., 2, 2, 9–60. PAVLOVIĆ, P.S. (1893): Prinove geološkog zavoda [New specimen of Geological Sur­ vey – in Serbian].– Ann. Geol. Pennins. Balk., 4, 186–191. PAVLOVIĆ, P.S. (1922): Prilozi za poznavanje tercijara u Srbij [Contribution to the study Tertiary in Serbia – in Serbian].– Ann. Geol. Pennins. Balk., 7/1 42–52. PAVLOVIĆ, P.S. (1925): O povredama na drugomediteranskim mekušcima iz Srbije [Breaches of the Second Mediteranian molluscs from Serbia – in Serbian].– Ann. Geol. Pennins. Balk. Zapisnici Srpskog geološkog društva, zbor 203, 18/1, 204–205. PAVLOVIĆ, P.S. & PETKOVIĆ, K.V. (1903): Prinove geološkog zavoda [New speci­ men of Geological Survey – in Serbian].– Ann. Geol. Pennins. Balk., 6/1, 293–325. PETROVIĆ, M. (1985): Foraminiferske zone i loze badenskog kata okoline Beograda [Badenian foraminiferal zones and subzones in Belgrade surroundings – in Ser- bian, Abstract in English].– Ann. Geol. Pennins. Balk., 49, 271–291. PETROVIĆ, M. & ŠUMAR, M. (1990): Biostratigrafski pregled sarmata okoline Be- ograda [Biostratigraphical overview of the Sarmatian in Belgrade surrounndings – in Serbian].– 12th Congress of the geol. of Yugoslia, Ohrid, 1, 162–173. PREACHER, K.J. (2001): April: Calculation for the Chi-square Test: An Interactive Cal- culation Tool for Chi-square Tests of Goodness of Fit and Independence [Compu- ter software].– Available from http://quantpsy.org. RŐGL, F. (1998): Palaeogeographic Considerations for Mediterranean and Paratethys Seaways (Oligocene to Miocene).– Annalen des Naturhistorischen Museums in Wien, 99A, 279–310. RUNDIĆ, LJ. (1992): Badenian ostracodes from Gornja Trnava area, NE Bosnia (in Ser- bian and in English).– Ann. Geol. Pennins. Balk., 56/1, 251–262. SAWYER, J.A. & ZUSCHIN, M. (2011): Drilling Predation in Mollusks from the Lower and Middle Miocene of the Central Paratethys.– Palaios, 26, 284–297, doi: 10.2110/ palo.2009.p09-161r SCHWARZHANS, W., BRADIĆ, K. & RUNDIĆ, LJ. (2015): Fish-otoliths from the marine-brackish water transition of Middle Miocene strata of Belgrade, Serbia.– Paläontologische Zeitschrift, 272, doi: 10.1007/s12542-015-0272-6 STEVANOVIĆ, P. (1951): Tragovi Panonskog mora u našoj zemlji [Traces of the Pan­ nonian Sea in our country – in Serbian].– Protection of Nature, 2–3, 5–26. STEVANOVIĆ, P. (1970): Paleogeografsko- ekološke prinove iz tortona okoline Be- ograda [Paleogeographic and ecological additions from Tortonian near Belgrade – in Serbian].– Glas SANU, 258, Odelj. Prir. mat. nauk., 33, 1–27. STEVANOVIĆ, P. & STEPANOVIĆ, B. (1939): Prethodna beleška o drugomediteranskim slojevima na levoj obali Topčiderske reke (Rakovica) i njihove faune [Previous record of the Second Mediteranian layers on the left bank of the Topčider river (Ra­ kovica) and their fauna – in Serbian].– Ann. Geol. Pennins. Balk., 16, 134–136. SZCZECHURA, J. & ABD-ELSHAFY, E. (1988): Ostracodes and foraminifera from? Middle Miocene of the western coast of the Gulf of Suez, Egypt.– Acta Palaeon- tologica Polonica, 33/4, 273–342. TOMAŠOVYCH, A. & ZUSCHIN, M. (2009): Variation in Brachiopod Preservation along a Carbonate Shell-basin Transact (Red Sea and Gulf of Aden): Environmen- tal Sensitivity of Taphofacies.– Palaios, 24, 697–716, doi: 10.2110/palo.2009.p09- 018r VERMEIJ, G.J. (1982): Gastropod shell form, repair and breakage in relation to preda- tion by the crap Calappa.– Malacologica, 23, 329–334. VERMEIJ, G.J. (1987): Evolution and Escalation: an Ecological History of Life.– Prin- ceton University Press, Princeton, New Jersey, 527 p. VERMEIJ, G.J. (2002): Evolution in the Consumer Age: Predators in the History of Life.– In: KOWALEWSKI, M. & KELLEY, P.H. (eds.): The fossil Record of predation. Paleontological Special Papers, 8, Yale University, New Haven, 375–394. VERMEIJ, G.J. & DUDLEY, E.C. (1982): Shell repair and drilling in some gastropods from the Ripley Formation (Upper Cretaceous) of the U.S.A.– Cretaceous Re- search 3, 397–403. doi: 10.1016/0195-6671(82)90041-6 VERMEIJ, G.J. & ROOPNARINE, P.D. (2001): Edge-drilling History and Distribution of a Novel Method of Predation.– PaleoBios, 21, Supplement to number 2, 130. WALKER, S.E. (2001): Paleoecology of Gastropods Preserved in Turbiditic Slope De- posits from the Upper Pliocene of Ecuador.– Palaeogeography, Palaeoclimatology, Palaeocology, 166, 141–163. doi: 10.1016/S0031-0182(00)00206-6 ZŁOTNIK, M. (2001): Size-related changes in predatory behaviour of naticid gastropods from the Middle Miocene Korytnica Clays, Poland.– Acta Palaeontologica Po- lonica, 46, 87–97. ZORN, I. (2003): Ostracoda from the Gaindorf Formation (Middle Miocene, Lower Ba- de nian) of Mühlbach (Molasse Basin, Lower Austria).– Annalen des Naturhistor- ischen Museum in Wien 104 A, 77–84.