A cave response to environmental changes in the Late Pleistocene : a study of Budimirica Cave sediments , Macedonia

Budimirica Cave is a small cave located in the southern part of the Republic of Macedonia, in the Kamenica Valley, a tributary to the Crna Reka and part of the Vardar river drainage network. The response of the cave to Late Pleistocene environmental changes is interpreted based on a detailed study of cave sediments, with previous data being supplemented, reinterpreted and compared to the Ohrid Lake palaeoclimate record. The oldest exposed speleothems in the Budimirica sediment profile were deposited during the MIS 5a (radiometric age of ca 83 ka). Sands to clays in the overburden are characterized by cycles separated by short-lived interruptions in deposition. They were deposited from a number of repeated flood events likely during the MIS 4 stage, the Weichselian (Wurmian) Glaciation, correlating to aggradation in the Kamenica Valley. The top flowstone is correlated with a warmer climate excursion at 45–50 ka (MIS 3) recor­ ded from the Ohrid Lake deposits. The whole section was extensively eroded during the MIS 2 stage, due to strong incision in the Kamenica Valley, indicated by knickpoint retreat. The erosion surface is overlain by a fossil-bearing breccia (with Ursus spelaeus) derived from frost shattering of the cave walls close to the entrance due to climate deterioration and enlargement of the entrance by slope retreat during the MIS 2 stage. Budimirica Cave sediments reflect changes in the Kamenica Valley, as well as the environmental changes during the last glacial-interglacial cycle, with clastic cave sediments deposited during glacial stadials, and erosion and flowstone deposition characteristic of the interstadials. They also allow reconstruction of the evolution of the Kamenica Valley during the Late Pleistocene, with a general trend of valley incision hindered by climate influenced river aggradation, but reinforced by river knickpoint retreat.


INTRODUCTION
Caves serve as especially good archives of past environmental information due to both the sensitivity of karst systems to a wide range of environmental changes (i.e.change of recharge dyna mics, increase/decrease and character of sediment supply, change of dissolved CO 2 content etc.), and great capability to protect the record from erosion as compared with the surface (e. g., HÄUSEL MANN et al., 2008MANN et al., , 2015)).Most of the recent studies have con centrated on analysis of palaeoenvironmental records from spe leothems (e. g., LASCU & FEINBERG, 2011;or summary in FAIRCHILD & BAKER, 2012).There exist only rare examples, when speleothems record longer timespan than relatively short climatic optima (like in Pleistocene), but such profiles are mostly prePleistocene in age (BOSÁK et al., 2002;HÄUSELMANN et al., 2015).Clastic cave fill preserves the palaeoenvironmental re cord and evidence of its changes in similar detail (see summaries a. o. in SASOWSKY & MYLROIE, 2004;FORD & WILLIAMS, 2007, or ZUPAN HAJNA et al., 2008) but usually over much longer timespans than speleothems.The disadvantage of such a record results from complications by the number of postdeposi tional changes (BULL, 1981) and hiatuses of variable lengths, sediment reworking and redepositon (BOSÁK et. al., 2000;ZU PAN HAJNA et al., 2008, 2010).Moreover, most caves preserve results of only the last depositional cycle, as deposits of previous cycles were usually exhumed due to the general function of cave systems.This can lead to the absence of cave sediments from certain timeperiods in whole regions, e.g.Middle Pleistocene sediments are missing in caves of the Bohemian Karst (central Czech Republic;HORÁČEK & LOŽEK, 1988).
We present an example of palaeoenvironmental analysis of clastic cave sediments belonging to the last depositional cycle of a small cave in the southern parts of the Republic of Macedonia.The interpretation is based on detailed research on the stratigra phy, composition and age of cave sediments.Previous data (TE MOVSKI, 2016) is supplemented, reinterpreted and compared to the palaeoclimate record from the Ohrid Lake sediments.The results provide insight into the effect of Late Pleistocene environ mental changes on landscape evolution in the area.

CAVE AND SEDIMENTS
The Budimirica Cave (400 m a. s. l.) is located in the right slope of the lower part of the Kamenica Valley (Figs. 1 and 2).The Ka menica River is a tributary to the Crna Reka, the second longest tributary of the Vardar River, the main drainage system of the Republic of Macedonia, draining towards the Aegean Sea.The area is a part of the Vitačevo Plateau (800-1.000m a.s.l.), com posed mostly of Pliocene and Early Pleistocene pyroclastic sedi ments.They were deposited in lacustrine to fluvial environments of the Tikveš Basin, as part of the PliocenePleistocene Central Macedonian Lake.In places they cover Palaeogene conglome rates (Tertiary Tikveš Basin), and rocks of the Vardar Zone pre Cenozoic basement: Cretaceous limestones and flysch (Figs. 1   , 1976;HRISTOV et al., 1965;RAKIĆEVIĆ et al., 1965;and RAKIĆEVIĆ & PENDŽERKOV SKI, 1970).and 2; DUMURDŽANOV et al., 1976DUMURDŽANOV et al., , 2004)).The Central Mace donian Lake was drained probably no later than during the Mid dle Pleistocene (DUMURDŽANOV et al., 2005) as a result of uplift in the central parts of the Balkan Peninsula and subsidence in the area of Aegean Sea (DUMURDŽANOV et al., 2004).The Kamenica River cut its valley through the pyroclastic deposits down to the preCenozoic basement, uncovering Upper Creta ceous (Turonian) limestones and exposing them to weathering and karstification (TEMOVSKI, 2016).
The Budimirica Cave is 105 m long and consists of two branching passages with northeastern general direction (MANA KOVIK, 1971).The cave passages formed along the strike direc tion of Turonian limestones and contain speleogens indicating spe leogenesis in phreatic to epiphreatic conditions (TEMOVSKI, 2016).A paragenetic morphology (ceiling channels, pendants) is developed especially above still preserved cave sediments (Fig. 4).
A large excavated pit in yellow to orangecoloured cave sediments (5 m long, 1.5 m wide, and 1-1.5 m deep) is situated in  the middle part of the Left Passage (Figs. 3 and 4).In the NE part of the pit, the sediment profile starts with flowstone at the bottom, which is followed by ca 1.5 m thick deposits of silt, clay and sand, covered by a 30 cm thick collapsed flowstone block (Figs. 4 and  5).Above (and beside) this flowstone block, in one part, the clas tic sequence continues with yellowish gravel, sand and silt up to 1 m thick, covered by flowstone (Figs. 5).This whole section is eroded and covered with brown fossilbearing breccia in the SW part of the pit (Figs.3C and 5).The breccia deposits are covered with a 10-15 cm thick organic deposit close to the cave entrance.
The clastic sequence below the collapsed flowstone block (0 cm), down to the bottom flowstone (155 cm) was studied in detail (TEMOVSKI, 2016).The sequence contains some principle un conformities expressed by dark films (Mn and/or Fe compounds) and desiccation cracks filled by overlying sediments, together with smaller neptunian dykes (Fig. 3B).Palaeomagnetic param eters of 56 samples collected from the middle part (between 25.5 and 155 cm) showed normal polarized magnetization with a mean palaeomagnetic declination of 2.6° (TEMOVSKI, 2016).

Palaeomagnetic analysis
Samples were analyzed at the Institute of Geology of the Czech Academy of Sciences.In total 20 oriented samples from previ ously non analyzed upper (clastic) and lower (flowstone) parts of the profile (Figs. 4 and 6) were collected for palaeomagnetic dat ing using the highresolution sampling methodology described by ZUPAN HAJNA et al. (2008).Palaeomagnetic data for the middle part of the profile was previously published, with 56 mea sured samples (TEMOVSKI, 2016).Unconsolidated sediments (13 samples) from the upper part of the profile (0-25.5 cm) were collected in small plastic cubes from individual horizons.Two solid hand samples from the speleothem in the lower part of the profile (168-181 cm) were cut into 7 cubes 20 x 20 x 20 mm.In the laboratory, they were measured on the JR6A spinner mag netometer (JELÍNEK, 1966) and/or 2G Superconducting Rock Magnetometer with incorporated AF unit.17 specimens were demagnetized by the alternating field (AF) procedures, up to the field of 100 mT in 10 to 13 steps and 3 specimens were demag netized thermally (TD) using the MAVACS apparatus (PŘÍHODA et al., 1989).The AF demagnetization by LDA-3A or SRM de magnetizer including TD demagnetization gave considerably good results.The remanent magnetization (RM) of specimens in their natural state (NRM) is identified by the symbol M. Graphs of normalized values of M/M o = F(H) were constructed for each analyzed specimen.Volume magnetic susceptibility (MS) was measured on a KLY4 kappabridge (JELÍNEK, 1973).Separa tion of the respective remanent magnetization components was carried out by multicomponent Kirschvink analysis (KIRSCH VINK, 1980).The statistics of FISHER (1953) were employed for calculation of mean directions of the pertinent remanence com ponents derived by the multicomponent analysis.Both previ ously published (56 samples) and new data (20 samples) were used for the statistical analyses, with data from four samples ex cluded where the maximum angular deviation (MAD) was higher than 10 for the multicomponent analysis, leaving a total of 72 analyzed samples.
Examples of the TD and AF demagnetization of samples (position 168 and 181 cm) with normal (N) palaeomagnetic po larity are presented in Fig. 6.Directions of Ccomponents of re manence with normal palaeomagnetic polarity for the whole pro file are given on Fig. 7. Recalculated mean palaeomagnetic directions of N polarized C-components for the whole profile are D = 359.0°and I = 46.5°.
The sediments from the profile showed only normal polarized magnetization.Mean values of palaeomagnetic directions of the whole profile are documented in Table 1.The values of palaeomagnetic directions in different segments document palaeo secular variation.The distribution of the NRM and MS in the pa laeomagnetic profile (Fig. 8) can indicate changes in deposi tion as well as the change of external climatic conditions.

Radiometric dating (U-series)
The Th/U (Useries) dating was performed in the Geochrono logy Laboratory, Institute of Geological Sciences, Polish Acade my of Sciences in Warsaw (analysts Helena Hercman).Standard chemical procedures for uranium and thorium separation from carbonate samples were used (IVANOVICH & HARMON, 1992).The activity was measured by α-spectrometry, using an ORTEC OCTETE PC.Spectral analyses and age calculations were made with URANOTHOR 2.5 software, which is the standard software in the Geochronology Laboratory in Warsaw (GORKA & HERCMAN, 2002).The quoted errors are one standard deviation.The bottom flowstone deposit (sample BUD_00) was dated to 83 (+16/14) ka (Fig. 5, Tab. 2).Dating of sample BUD_02 (which belongs to a stratigraphically older layer) was not successful due to strong detrital Th contami nation.

Ohrid Lake palaeoenvironmental record
Detailed studies of sediments from Ohrid Lake (Fig. 1), from seve ral cores at different locations, have been undertaken recently, with published data providing palaeoenvironmental information dating back to the Middle Pleistocene (LACEY et al., 2016 and references therein).
To compare with the environmental changes registered within the Budimirica Cave sediments we used Late Pleistocene palaeoclimate data derived from a study of a multiproxy record from the sediments of JO20041 core, from the southwestern part of Ohrid Lake (BELMECHERI et al., 2010;LEZINE et al., 2010).

Cave response to Late Pleistocene environmental changes
Passages of the Budimirica Cave formed along the strike direc of the Budimirica Cave formed along the strike direc the strike direc tion of the Turonian limestones in phreatic to epiphreatic environ of the Turonian limestones in phreatic to epiphreatic environ ments (Fig. 9a).The age of speleogenesis cannot be precisely stated, but it is possible that it was connected with the period be fore the incision of the Kamenica River Valley, indicated by small cave remnant on the opposite side of the valley (the Karši Budi mirica Cave; TEMOVSKI, 2016).Paragenetic speleogens (ceiling channels, pendants) are developed in many places, especially above the preserved cave sediments, indicating the upward de velopment of passages after sediment aggradation in the vadose evolution stages of the cave (TEMOVSKI, 2016).
The flowstones on the passage floor were deposited in va dose conditions (Fig. 9b).A radiometric age of ca 83 ka suggests an origin during the interstadial of the Marine Isotopic Stage (MIS) 5a.This period was characterized by a warm and wet cli mate, supporting an increase in vegetation cover and carbon di oxide production according to the Ohrid Lake record (BEL MECHERI et al., 2010;LEZINE et al., 2010), although having less favourable conditions compared to the Last (Eemian) Inter glacial (MIS 5e).The dated flowstone layer (sample BUD_00) represents the upper layer below the clastic deposits, and it is covering an older layer of flowstone separated with a thin clay layer.Although the dating of the older layer was unsuccessful, the stratigraphy (Fig. 5) indicates that the lower layer (BUD_02) was likely deposited in older interstadials during MIS 5, with clay deposits deposited during the stadial period.
The basal flowstone is covered with clastic deposits (up to the collapsed flowstone block) composed of two different lithos tratigraphic segments divided by a distinct unconformity at 123 cm: a sandy one in the lower part and a generally fine-grained one above (TEMOVSKI, 2016).The sandy character of the lower segment indicates the higher energy of the flowing water than in later overlying cycles.Upper segment sediments are composed of well-sorted fine-grained sediments, and are a typical example of internal cave facies lithology (sensu KUKLA & LOŽEK, 1958).Sediments were sorted when transported through the shal low phreatic/epiphreatic siphonal cave system, where the coarsergrained components were deposited closer to ponor(s).The cave fill represents the results of repeating floods in a rela tively calm but not completely stagnant sedimentary environ ment (slackwater deposits sensu BOSCH &WHITE, 2004), i. e. originating from waning floodwaters or other pulsed flow (cf.FORD & WILLIAMS, 2007).Such a sedimentary environment has been known from number of caves in the north Dinaric Classical Karst and adjacent areas (e.g., ZUPAN HAJNA et al., 2008, 2010).
The profile consists only of a few single flood events (cycles) limited by unconformities with shortlasting hiatuses (compac tion and dewatering, desiccation, small erosions).Each flood partly eroded older sediments and deposited its load  The studied sediments contain minerals that are also typical for the volcanic and volcaniclastic source rocks and their weathe ring products -especially sanidine, smectite (motmorillonite) and kaolinite (TEMOVSKI, 2016).According to the kaolinite content and absence of macroscopically visible sedimentary particles de rived from volcanic and volcaniclastic rocks, the source material was composed of redeposited materials derived from wellde veloped weathering profile(s).
Only normal polarized magnetization is typical for the stud ied sediments.Mean values of palaeomagnetic directions calcu lated from the whole profile are close to the present magnetic field -the value of mean palaeomagnetic declination of 359° is close to the present magnetic declination for Macedonia with a value of 2.7°.Sediments are not older than 780 ka (Brunhes chron) as also indicated by the radiometric age of the underlying flowstone.The distribution of the MS on the studied profile can indicate changes in deposition as well as the change of external climatic conditions.
Considering that the clastic sequence follows the MIS 5a flowstone, it is assumed to be deposited during the MIS 4 stadial (Fig. 9c).Lacustrine deposits of that age from the Ohrid Lake (BELMECHERI et al., 2010;LEZINE et al., 2010) record a gene rally cold and dry climate.The Budimirica fill indicates chang ing climatic conditions with several relatively shortlasting but quite massive floods reflecting most probably highly fluctuating periodic precipitation rates.
The deposits are eroded below the paragenetic ceiling and covered with the top flowstone, indicating another change of the depositional environment, towards a rather slower lowvolume water flow (Fig. 9d).The top flowstone was likely deposited dur ing the MIS 3 interstadial, as it follows the MIS 4 clastic se quence and is followed by strong erosion overlain with MIS 2 deposits.The Ohrid Lake sediments (BELMECHERI et al., 2009(BELMECHERI et al., , 2010) ) registered the MIS 3 warmer and wetter climate ex cursion at 45-50 ka.
The whole flowstone-clastic sediments-flowstone sequence was then deeply eroded and the cave fill was exhumed extensively (Fig. 9e), likely by the invasion(s) of allogenic waters triggered by the strong incision of the Kamenica Valley, continuing water table fall, and change of environmental conditions during the MIS 2 stage.
The deep erosion of both clastic sediments and the topmost flowstone was followed by the deposition of fossil-bearing brec cia close to the cave entrance (Fig. 9f).Angular limestone frag ments in the breccia deposit represent the products of frost shat tering of cave walls especially at and near the cave entrance.This can be a result of both climate deterioration and enlargement of the cave entrance due to slope retreat in the valley, which corre lates well with the strong incision of the Kamenica Valley indi cated by the previous erosion of cave sediments.Among several other bones, two teeth belong to Ursus spelaeus (R. GAREVSKI, pers. comm., 2012;J. WAGNER, pers. comm., 2013), which can be compared with late Weichselian (Wurmian 2/3 and Wurmian 3) bear remains from the Makarovec Cave located in the Babuna Valley in a neighbouring basin to the north of the studied cave (Fig. 1); GAREVSKI (1969) and GAREVSKI & MALEZ (1984).Therefore the breccia can be attributed to the latest cold period of the Weichselian glaciation (MIS 2).Ohrid Lake sediments in dicate unstable environmental conditions during this period with a fluctuation of precipitation and temperature (LEZINE et al., 2010).
The top organogenic deposits are likely due to recent (Holocene) pasturage and represent sheep and/or cow coprolites (Fig. 9g).

Valley incision and aggradation
The clastic deposition (as well as associated paragenetic morpho logy) in Budimirica Cave correlates with the remnant of a river terrace located in the gorgelike Kamenica Valley opposite the cave (Fig. 10 increases the erosion potential in the basin.This was likely rein forced by the seasonality of precipitation and runoff, thus increas ing sediment productivity and forcing aggradation of the river valley.Such conditions (enhanced erosion) were also registered in the drainage basin of the Ohrid Lake during the Late Pleis tocene cold stadials (LEZINE et al., 2010).
The longitudinal profile along the Kamenica Valley shows several knickpoints (sharp changes in channel slope; Fig. 10).One is located a little upstream from Temna Peštera -Dragožel, at the contact of the limestone and flysch, and is fixed (structurally and hydrologically), with the upstream karst system discharging along the contact and providing an increase in river potential for erosion.Another (fixed) one is downstream from the Budimirica Cave, along the fault at the contact of the flysch and limestone, and may also indicate uplift of the limestone block.A third knick point can be noticed a little upstream from the limestone block where the Budimirica Cave is developed.The location of this knickpoint (not at the contact of limestone and flysch, but further upstream, in the flysch) and the higher river gradient in the lime stone block indicates headward retreat of the knickpoint, and thus stronger downstream erosion, which could be responsible for the strong erosion registered in the Budimirica Cave during the MIS 2.
Such strong erosion is also indicated by the average incision rate of the Kamenica Valley based on the radiometric ages of the flowstone deposits in the Budimirica Cave and Temna Peštera -Dragožel (Fig. 10).The bottom flowstone in the Budimirica Cave sediment profile (dated at ca 83 ka) was deposited when the water table was a few metres lower than the cave (~395 m), and thus with relative elevation of 80 m above the Kamenica River, this gives an average incision rate of ~0.96 mm.y 1 for the last 83 ka.In the middle part of the Kamenica Valley, at Temna Peštera -Dragožel, the upper flowstone (covering the clastic deposits), which is at least 350 ka old (and follows previous aggradation), points to (a maximum) average incision rate of 0.23 mm.y 1 .This value is almost five times smaller than that below the Budimirica Cave, although it represents a longer period (which could also in clude longer periods of incision paucity and/or aggradations).However, the incision in the lower part is still quite high, and cor relates well with the strong erosion registered in the Budimirica Cave sediments.

CONCLUSION
Cave deposits preserved in the Budimirica Cave reflect the re sponse of the cave system to the environmental changes in the drainage basin.The Budimirica Cave is part of a contact karst system receiving both allogenic and autogenic recharge.Incision of the Kamenica Valley lowered the water table and allowed de position of flowstone under vadose conditions during the MIS 5a.Flowstone deposits were covered by allogenic clastic sediments, originating from the weathered profiles of pyroclastic sediments of the Plio-Pleistocene Vitačevo and Mariovo Formations.They were deposited in the cave in several floods as a result of the in crease of sediment supply and a base level rise due to aggradation in the Kamenica Valley in the MIS 4, which also forced upward paragenetic development of the cave.Continuing incision of the Kamenica Valley allowed erosion of the upper part of the clastic sequence and deposition of the upper flowstone deposits during the MIS 3, followed by strong erosion during MIS 2 as a result of fast incision triggered by river knickpoint retreat.This influenced enlargement of the entrance by slope retreat and deposition of fossil bearing breccia due to frost shattering during the cold sta dial.Recent Holocene input is restricted to human pasturage in fluence with deposition of organogenic deposits.
The interpreted evolution of the cave deposits indicates that flowstone deposition corresponds to glacial interstadials (MIS 5a and MIS 3), and clastic sediment deposition corresponds to gla cial stadials (MIS 4 and MIS 2).Cold periods affect vegetation cover and soil activity, which together with the seasonality of pre cipitation increases erosion in the basin, and the increase of sedi ment productivity forces aggradation in the valley, which also affected the cave development.The warm and wet climate of the interstadials, supports vegetation productivity and soil develop ment, which reduces erosion and sediment supply, and allows speleothem deposition in the cave.
The general trend of incision in the Kamenica valley was hindered due to climate induced aggradation during the MIS 4, but continued during MIS 3 and was reinforced by knickpoint retreat during MIS 2. This change also affected the cave, with incision lowering the water table and influencing erosion of the cave sediments.
Dating of the changes in the Kamenica Valley registered in cave sediments from the Budimirica Cave gives information about valley incision rates along the Kamenica Valley, with high incision rates in the lower part of the Kamenica Valley being in agreement with the inferred incision rates of Quaternary valleys in the mountainous areas of Macedonia (MILEVSKI & KOLČA-KOLČA KOVSKI, 2012).

Figure 1 .
Figure 1.Location of the studi ed site within the regional geo logical structure of the Republic of Macedonia.Major geo tec tonic units are after ARSOVSKI (1997).

Figure 3 .
Figure 3.View of the studied sediments in Budimirica Cave: A -The whole profile below the collapsed flowstone block with major unconformities indicated; B -Closeview of the upper part with positions of palaeomagnetic sampling boxes and examples of neptunian dykes (n) along the plane at 25.5 cm; C -Relation ship of yellow sediments to the brown breccia deposit, with the black ellipse showing the location of collected fossil samples and closeview photo showing frag ments of an Ursus spelaeus molar.Photos by M. Temovski.

Figure 4 .
Figure 4. Detail of the Left Passage with cave morphology, sediment relationships and location of the sampling profile.Modified from TEMOVSKI (2016).

Figure 5 .
Figure 5. Relationship of sampled sediments to other cave fills and the position of major unconformities (dotted line) in the pit of the Left Passage.Upper illustra tion (A) shows a frontal view of the lower part of the sampled profile with locations of the flowstone samples.Modified from TEMOVSKI (2016).
A longitudinal profile of the Kamenica River was constructed from 1:25000 topographic maps (VOJNOGEOGRAFSKI INSTI TUT, 1973) using the Global Mapper v.16 and Grapher v.11, and was analyzed to identify changes in channel slope, with geologi cal information along the profile obtained from 1:25000 geological maps (GEOLOŠKI ZAVOD -SKOPJE, unpublished).Field observations in the Kamenica Valley near Budimirica Cave registered the remnant of a river terrace opposite the cave at 80 m relative elevation.Combined field and topographic map observations registered three river terraces in the middle part of the Kamenica Valley, nearby Temna Peštera -Dragožel, with relative elevations of 30 m, 60 m, and 80 m above the Kamenica River.Their position shows their relationship to caves in both lo calities (TEMOVSKI, 2016).

Figure 6 .
Figure 6.Example (A) of the AF demagnetization (position 181 cm) and (B) the TD demagnetization (position 168 cm) of samples with normal palaeomagnetic po larity from the Budimirica Cave.Upper left -a stereographic projection of the natural remanent magnetization of a sample in the natural state (cross section) and after progressive AF and/or TD demagnetization; upper right -the Zijderveld diagram: solid circles represent projection on the horizontal plane (XY), open circles represent projections on the north-south vertical plane (XZ); lower left -a graph of normalized values of the remanent magnetic moments versus demagnetizing fields: M -modulus of the remanent magnetic moment of a sample subjected to AF and/or TD demagnetization.Lower right -a graph of the normalized values of volume magnetic susceptibility versus thermal demagnetizing fields: k -value of volume magnetic susceptibility of a sample subjected to TD demagnetization.

Figure 7 .
Figure 7. Directions of Ccomponents of remanence with normal palaeomag netic polarity for the whole profile in Budimirica Cave.Stereographic projection, full (open) small circles represent projection onto the lower (upper) hemisphere.The mean directions calculated according toFISHER (1953) are marked by a small cross across a circle, the confidence circle at the 95% probability le vel is circumscribed about the mean direction.

Figure 8 .
Figure 8. Basic magnetic properties for the whole profile in the Budimirica Cave.Data in red is from this study, others after TEMOVSKI (2016).Note that flowstone samples BU01 (AD) are younger than samples BU2 (AC), and are found at lower position as a results of the slope of the flowstone deposit (see Fig. 5 for clear view of relationship).
. The depo sitional and postdepositional textures (e.g.waterescape struc tures) indicate the very rapid depositional rates, which is sup ported by uncentered secular variation detected from the calculation of mean values of palaeomagnetic directions in the upper profile segment.The cyclic character resulted from a fluc tuating hydrological regime in the cave and reflected environ mental changes influencing the recharge regime and base level position in the Kamenica Valley.Individual layers, which also differ in their magnetic properties (MS and NRM values), reacted to changes in the catchment area (change of source rocks, con tinuing erosion of weathering profiles) and external climatic changes (cf.e. g., SROUBEK et al., 2001).

Figure 9 .
Figure 9. Schematic illustration of the evolution of Budimirica Cave deposits and their relationship to the changes in the Kamenica Valley with their interpreted ages.See text for explanation.Red line represents changes in elevation of Kamenica River talweg.

Figure 10 .
Figure 10.Longitudinal profile along the lower and middle part of the Kamenica Valley, showing locations of caves, river terraces, knickpoint positions and inferred valley incision rates.

Table 1 .
Mean palaeomagnetic directions for the whole profile in Budimirica Cave, recalculated using data from this study and TEMOVSKI (2016).