RMZ - Materials and Geoenvironment, Vol. 49, No.1. pp. 61-84, 2002 Where does Underground Ljubljanica Flow? Kje teèe podzemska Ljubljanica? FRANCE UTERIÈ1 Department of Geology, University of Ljubljana, Akerèeva 12, 1000 Ljubljana, Slovenia; E-mail: france.sustersic@ntfgeo.uni-lj.si 1 Received: January 29, 2002 Accepted: February 02, 2002 Abstract: The paper presents a field example of determination of the spatial position of the flow corridors within the krast, based on the data collected during when mapping denuded cave features (surface caving). The sediments within denuded caves provide a means of recognizing individual cave system sectors that display different histories of sedimentation, outwash and resedimentation, as well as possibly different sediment source areas. Such observations provide information about the spatial relationships of karst channels within a flow corridor. Reconstructed flow corridors north-east from Planinsko polje fit well to the results of extensive water tracing in the past. Izvleèek: Predstavljeni so terenski rezultati ugotavljanja prostorskega poloaja glavnih pretoènih koridorjev v krasu s pomoèjo podatkov, zbranih pri kartiranju denudiranih jam (surface caving). Sedimenti v denudiranih (brezstropih) jamah omogoèajo, da razloèimo posamezne dele pretoènega koridorja, ki so preiveli razlièno zgodovino sedimentacije, spiranja in ponovnega zatrpavanja, lahko tudi iz drugaènega izvornega obmoèja. Tako dobimo informacijo o prostorskih odnosih med krakimi kanali znotraj pretoènega koridorja. Rekonstruirani pretoèni koridorji severovzhodno od Planinskega polja se dobro ujemajo s povezavami, ugotovljenimi z neoposrednim slednjem podzemskih vod. Key words: karst surface, karst hydrology, speleogenesis, karst of Slovenia, cave sediments, surface carving Kljuène besede: krako povrje, hidrologija krasa, speleogeneza, kras Slovenije, jamski sedimenti, brezstropne jame Scientific Paper Izvirni znanstveni èlanek 61 62 INTRODUCTION The Ljubljanica river (sensu lato) is one of the best known karst sinking rivers, as it disappears/reappears seven times (and, thus, carries seven names), before it becomes a normal river, crossing Ljubljana, the capital city of Slovenia. Only two (out of four) of the main groups of karst springs that contribute to the river at Vrhnika are popularly named the Ljubljanica (sensu stricto). The previous extensive underground water tracing experiment (GOSPODARIÈ AND H ABIÈ , 1976) revealed that, though all of them share the same general catchment area, functional similarities would more reasonably suggest three groups. However, in more general discussions such as this one, they may all be considered together as the Ljubljanica springs. Before reaching the final springs, most of the water feeding the Ljubljanica springs crosses at least some of the karst poljes. UTERIÈ (1996) added support to GAMS (1959) idea that the underground connections between poljes with partly non-karstic inflow and the downstream poljes are, to a great extent, epiphreatic caves that are at least theoretically accessible to Man. On the other hand, the present outflows from the poljes with completely karstic inflow are deep phreatic, and have remained mostly unexplored by cavers. This is best expressed in the area between Planinsko polje (the lowest in the string) and the Ljubljanica (sensu stricto) springs at Vrhnika. UTERIÈ, F. Though complicated and varying due to regional oscillations in precipitation (input), water connections have been studied extensively, and have been well-established in the past (GOSPODARIÈ AND HABIÈ, 1976). Since the work of PUTICK (1887), several patterns1 of possible underground connections have been proposed. Among these, those of MICHLER (1954/55), and KRIVIC, VERBOVEK AND D ROBNE (1976) 2 are the best known (Figure 1). Application of surface caving as a method of systematic study of denuded, unroofed caves (MIHEVC, 1996; see also UTERIÈ, 1998, 1999) introduced a new set of data. Together with other minor indications, these observations permit recognition, at least to some extent, of the main flow corridor(s) in the area northeast of the Planinsko polje. This paper sets out to compile and to reconsider the existing data, and to reveal the relationships between the main stream corridor and well-marked tectonic zones. STUDY AREA AND GEOLOGICAL SETTING In this paper the main points of interest are the outflow part of Planinsko polje and Ravnik, a low-relief corridor about 18km long and 1 to 3km wide, lying about 3km northeast of the polje. 1 In his map Gradinica (Teufelshöhle = Devils Hole) is shown in the wrong position. 2 The map was intended to be Annex 2 of the book (GOSPODARIÈ & HABIÈ, 1976, ref. p. 213). After it was printed, the Yugoslavian military authorities declared it to be a military secret and did not permit its release to public. Nor has it been published since, so Fig.1 of this paper, though simplified and mutilated, is the first published version of this map. RMZ-M&G 2002, 49 Where does Underground Ljubljanica Flow? 63 Figure 1. Hydrogeological map of the treated area. Extracted from Hidrogeoloka karta 1: 50 000, appendix to GOSPODARIÈ & HABIÈ, 1976). Green - predominantly dolomite; blue predominantly limestone; yellow - alluvium, predominantly flooded bottoms of karst poljes; brown - nonkartsic rocks. Capital letters, added by the present author, mark main anticipated flow corridors (see text!) RMZ-M&G 2002, 49 64 Planinsko polje is one of the best-studied parts of the Slovenian Classical Karst and, thus, little background information need be repeated here. The reader is referred to works by ÈAR (1982), GAMS (1963), GOSPODARIÈ (1982-a, 1982-b), G OSPODARIÈ & H ABIÈ (1976), UTERIÈ and P UC (1970), and UTERIÈ (1982, 1996, 2000). However, to aid understanding of the remaining discussion, it is necessary to be aware that the sinking river is called the Unica, that the elevation of the polje flat is about 445m a.s.l., and the elevation of the main springs at Vrhnika is somewhat less than 300m a.s.l. Ravnik has been studied more recently (UTERIÈ, 1994, 1996, 1997, 1998, 2000). Earlier geomorphological research viewed Ravnik as a dry valley, presumably formed by the Cerknièica river (see UTERIÈ, 1996, Fig.2). This assumption was based only upon morphological similarities, without any material proof. It was not until the mid 1970s (UTERIÈ, 1976; MIHEVC, 1979) that conglomerates similar to the present Cerknièica river transport material were first found. In most outcrops the conglomerate appears together with flowstone, so that previous explanations were shaken rather than proved. Structurally, the northeastern part of Planinsko polje and Ravnik is a monocline, with beds dipping at 25o to 30o towards the westsouthwest. All of the rocks form part of the c.6,850m-thick Dinaric carbonate sequence. The oldest beds, in the southeast, comprise Late Triassic (Norian and Rhaetian) dolomites, whereas the youngest beds, in the northwest, are early Cretaceous limestones and dolomites, topped by nearly 100%-pure Late Cretaceous limestones. UTERIÈ (1996, 1998) showed that the contacts between lime- UTERIÈ, F. stone and dolomite packages within the Jurassic and early Cretaceous sequences are especially prone to cavern formation (PEZDIÈ, MIIÈ, UTERIÈ, 1998). The most important tectonic line in the area is the Idria Fault, running along the eastern margin of the Planinsko polje. In Ravnik two lineaments, respectively about 3km and 5km eastward from the polje (Figure 2, M, N), were until recently only partially recognised as major crush zones, and these run parallel to both sides of the Ravnik. Some displacement may have occurred, but it is insignificant if compared to the width of the zone(s). As demonstrated in the discussion below, these tectonic lines act as delimiters between zones of different cave sedimentation. Between the lineaments is a swarm of sinistral strike-slip faults, running northortheast, evidently having formed as a reaction to dextral strike-slip along the Idria fault. All of the large collapse dolines that have been studied in detail lie on such faults (see UTERIÈ, 1998). At first glance the string of currently inactive collapse dolines north of the village of Laze (Slaven dol string) (Figure 2, L) appears to be somewhat similar. UTERIÈ ET. AL. (2001) demonstrated its important role in the speleogenesis of the underlying cave system and argued that it deflected the main underground streams northwards, parellel to it. Current detaliled mapping of the wider area revealed that this structure is more complicated than previously expected and further research is in progress. RMZ-M&G 2002, 49 Where does Underground Ljubljanica Flow? 65 Figure 2. Main structural elements: red - Idria fault zone; yelow lineaments mentioned in the text (see explanation of letter symbols in the text). RMZ-M&G 2002, 49 66 HYDROGEOLOGY AND SPELEOLOGY OF THE OUTFLOW PART OF PLANINSKO POLJE General situation As is evident from Figure 1, an important collector of karst water (B) should run approximately in the direction of Ravnik. However, the map is essentially a topological view, merely indicating the mutual connections, whereas the actual spatial position of the individual links is purely conjectural. Nevertheless, the map (KRIVIC, VERBOVEK & D ROBNE , o.c.) clearly illustrates the authors idea that the three main north-directed flow corridors (Figure 1, A, B, C) must somehow be guided by the strike of the dolomite packs within the limestone. In addition to local precipitation water, input to the Ravnik system comes from two3 directions - from the southeast (i.e. from Cerkniko polje4 ) and directly from the west (i.e. from Planinsko polje). Ponors at Planinsko polje appear in distinct clusters, which carry their own names. For current purposes, a simple division into southern and northern groups will suffice (Figure 2, S, N / blue). Southern ponor group Most of the southern ponors lie along the eastern border of the polje. The southernmost UTERIÈ, F. in particular are fed permanently by sinking river water. According to the latest published data (BREZNIK, 1998, p.22) The average inflow into the polje is 26 m3s-1 The swallow holes can swallow a total of 60 m3s-1, those in Babni dol5 alone being capable of swallowing as much as 40 m3s-1. Subtraction yields 20 m3s-1 for the southern ponors when both groups are fully active. Their capacity is sufficient to keep the polje flat dry during most of the year. Vertical ponors dominate, whereas ponor caves are few and of short length. Even Jama na Lokah (cad. no.6 172), which for more than a century was estimated to be a few hundred metres long, turned out to be no longer than a few dozen metres (LAJOVIC, 2001). The eastern polje margin is quite straight, implying that it is guided tectonically. Detailed geological mapping at 1:5000 scale (ÈAR, 1982) revealed that the easternmost fracture of the Idria fault zone runs just a few dozen metres from the polje border, beneath the poljes alluvial flat. UTERIÈ & JAKOPIN (1979) demonstrated that most of the ponors lie at the foot of the steep polje wall, rather than in areas of gentler slope. This indicates (a) that there is strong tectonic guidance of the polje margin; (b) that the polje is relatively young, so that differential slope weathering has not yet been able to make it more lobate; and (c) that sinking water prefers less tectonised zones to more crushed ones. 3 Input from the northwest, in the area of Logatec, 6km north of Planinsko polje, is ignored in the general discusssion and mentioned only where needed. 4 Some quantity even directly from the canyon before the river reaches the polje basin. 5 In terms of the present paper approximating to the northern ponors. 6 According to the central cave register of Slovenia, maintained by the Karst Research Institute, Postojna, and Speleological Asociation of Slovenia, Ljubljana. RMZ-M&G 2002, 49 Where does Underground Ljubljanica Flow? Water tracing experiments in 1975 (GOSPODARIÈ AND HABIÈ, 1976) revealed that the highest velocities (c.6 cm s-1), within the karst underground between the polje and the springs, are between the southern ponors and the main springs at Vrhnika. As mentioned above, there is no ponor cave that would permit direct access into the karst underground farther from the polje. During the highest floods water appears in Maèkovica (cad. no. 52) and Logarèek (cad. no. 28), caves that have their own entrances in the slope above the polje. The latter is the larger of the two and, as both are part of the same system, the former, which lies closer to the polje, is generally omitted from discussions. The entrance shaft of Logarèek leads to an epiphreatic channel of moderate dimensions, at approximately 470m a.s.l., i.e. about 25m above the present polje floor. Another drop of about 25m makes more than 1km of passage accessible at the approximate level of the polje. This passage is somewhat larger than the higher one and, despite being modified by drip water, is basically tunnel-like, i.e. epiphreatic. In cavers jargon it is termed a main drain, and during the highest floods in the polje it becomes partly flooded. The bottom of the main drain is perforated by (presumably) vadose shafts, which become dry only during extreme draught. Recently, cavers discovered a continuation of the upper passage, which led them to a maze, as yet only partly explored, at the approximate level of the main passage. The main drain winds in a generally northward direction. Scuba diving revealed a 7 67 drowned passage only a few metres below the water table, continuing in the same general direction. The end part of the cave is accessible only in dry summers, and it was explored for just a few hundred metres before the divers turned back without reaching a physical end (PALÈIÈ, 2001). Presently, Logarèek is partly flooded (to elevations of 444 to 447m a.s.l.) only at the time of high water in the polje. GAMS (1963), who presented a comprehensive study of Logarèek, demonstrated that there is no third completely drowned discrete channel lower than the main channel, as originally postulated PUTICK (1887), but that several only partly interrelated streams must pass below the cave. Water tracing experiments in 1975 proved a relatively direct connection between the Dolenje Loke ponor (northern part of the southern group) and Logarèek (GOSPODARIÈ and HABIÈ, 1976, 1987 ). An early observation (August, 30, 1928 / cit. GAMS, o.c. p.55) that warm summer water somehow reaches the most distant parts of the cave, whereas colder, winter, water remains stored in water bodies close to the polje, was recently proved by divers (PALÈIÈ, o.c.). Loamy sediments (coarser-grained sediments are not known) and various flowstone formations in Logarèek, as well as other caves in the system, have not yet been studied. The existence of clearly epiphreatic, partly canyon-like passages, and the lack of coarser sediments, appear somewhat enigmatic. Fragments of higher, older, abandoned cave passages in the area in direct contact with Calculated velocity of 11.4 cm s-1 or 1,500m per 4 hrs. RMZ-M&G 2002, 49 68 the southern ponors are scarce. Unroofed caves have not been searched for systematically, but the general impression is that they are few. Northern ponor group Physically, the northern group of ponors begins only a few hundred metres farther to the northwest, along the polje margin. Effectively, water must flow a few kilometres on the surface around Jakovica hill (which takes the form of a peninsula during normal floods) before reaching them. At first glance the northern group is similar to the southern there are as vertical, as cave ponors. Again, the caves are short, though somewhat larger. They appear to narrow down, rather than being choked by collapse, and the general impression is that they are relatively young. The situation in two artificial ponorwells that penetrated natural caverns (PUTICK, 1889; GAMS, 1963) is little different. The northern ponors are active only at middle and high water levels; during the summer they generally remain dry. Nevertheless, since PUTICK (1887) they have been regarded as the main ponors of Planinsko polje. JENKO (1954, cit. GAMS, 1963, p.9) estimates that the southern ponors have a capacity of 17 m3s-1 during the low and middle water levels in the bed of the Unica, and they can swallow the whole river. During flood the swallow holes in the extreme northern corner of the polje, at the foot of Lanski vrh (hill), absorb 43 m3s-1, or in exceptional cases UTERIÈ, F. as much as 60 m 3s -1.(Translated by F. uteriè) The caves further from the polje margin differ significantly from those related to the southern sector. Beside the main one, Najdena jama (cad. no. 259), presently a 6,5 km-long phreatic maze, there are several dozen fragments with accessible lengths ranging from ten8 to a few hundred metres. Except for Vranja jama (cad. no. 88), which can be considered together with the nearby Najdena jama, all of them are dry, high above the karst water level undulations. Najdena jama is accessible through a presumably vadose shaft formed in massive flowstone, indicating that the surface has lowered significantly since its formation. An attempt to interpret the cave in terms of standard fluvial levels (GOSPODARIÈ, 1982-a), was not convincing (UTERIÈ, 1994). The so-called horizontal passages are in fact segments of a 3-D phreatic maze, parts of which appear horizontal due to abundant loam sedimentation. When the polje is dry, water levels in the siphon lakes range between 410m, a few hundred metres from the polje margin, and 396 m in the most distant parts. During floods, when the cave can be more than twothirds water-filled, the piezometric level in most of the cave stabilises at 435 m. Najdena jama is well known for its enormous quantities of mud, which testify the discharge of high water through it to be small, even 8 According to the rules of the Speleological Association of Slovenia caves shorter than 10m are not registered. In this context, the term cave is used in the strict cavers meaning: accessible natural void of prescribed dimensions. So, a short fragment of a large system is treated as a cave, in just the same way as a maze several kilometres long, if they have independent entrances and are not connected physically. Very recently, cave fragments, shorter then 10m, have been attributed to unroofed caves which have not yet been mapped as systematically as true caves in this area. One may expect that this group will soon outnumber the registered caves. RMZ-M&G 2002, 49 Where does Underground Ljubljanica Flow? though the passages are nearly completely water-filled during floods. At times of middle flow, when the Unica just reaches the two artificial ponor wells, a stream of about 1 m3s-1 appears in some passages, and these are the only really washed ones. Detailed monitoring of floods in Najdena jama and on the polje revealed that, when the water level within the cave starts to increase, the difference between the polje and cave water levels remains nearly constant (15m), regardless of the absolute level of the water on the polje. This was interpreted as indicating the presence of a strong filter between the polje and the cave (UTERIÈ, 1982, p.23). In such a situation, at least within Najdena jamas area of influence, which earlier researchers considered has the most effective swallow holes, a significant increase in swallow hole capacity due to increases of flood level on the polje would not be expected. Clastic sediments in the cave were studied extensively by GOSPODARIÈ (1982-b). He found that the oldest are early Rissian quartz sands and loams of unknown origin. More interesting is the gravel, including various types of chert. In Planinska jama, the main inflow cave of the Planinsko polje, distinct cherts were deposited separately during the Rissian (GOSPODARIÈ, 1976), whereas in the system of Najdena jama they were deposited, mixed together, at the beginning of Würmian. Presumably, rapid erosion in Planinska jama led to their retransportation to caevs on the northern side of the polje. Younger clastic sediments include various limestone gravels, and pebbles of local rock, 69 mostly derived from cryoclastic rock spalling during cold phases of the Würm. Input of this material resulted in canyonisation of the southern part of the Piparski rov 1 (UTERIÈ, 1991, p.81), which extended about 500m into the system, away from the polje, until the pebbles (the abrasive) were worn out. GOSPODARIÈ (o.c., pp.186-197) presumes that the oldest flowstone within the cave is of Riss/Würm age. An attempt to use the U-Th series method to date the massive flowstone around the present cave entrance failed, because the flowstone is too old (MIHEVC , 2000). Fragments of higher, older, abandoned cave passages in the area in direct contact with the northern ponors are quite abundant. Other than Skednena jama and the previously mentioned Piparski rov 1 in Najdena jama (which are the only locally conditioned exceptions), all of them are clearly of phreatic origin. It appears that they all formed within the same tier, with a vertical extent of at least 150m (UTERIÈ, 1994). Lithological guidance is evident, in the sense that the main inception horizons are the upper and, especially, the lower contacts of the dolomite packages9 . Relatively common unroofed caves also appear to be phreatic. The sediment fill ranges from coarse conglomerates of local rock, to various sands, loams, diverse cave originated septarian concretions, and abundant flowstone. Unfortunately, only pilot studies of the sediments have yet been carried out, confirming that an admixture of flysch sand, presumably originating from the Postojna area, is always present. 9 Interesting enough, first observed in Najdena jama by GOSPODARIÈ, 1982-a, 167, without an idea for further discussion. RMZ-M&G 2002, 49 70 About 3km northwards (i.e. downstream) from Najdena jama is the Gradinica cave. At the bottom of a quite large, 189 m-deep shaft is a chamber of 0.375 Mm3. Its floor lies at an elevation of 379m, which is 65 m below the level of Planinsko polje, and 30 m below the deepest, subaerial, part of Najdena jama. During floods, the rise of water level in Gradinica would not affect the water level in Najdena jama until an elevation of somewhat more than 400 m is achieved. Beyond this point both levels increase more or less in parallel. Sediments in Gradinica have not yet been studied. The only analyzed sample of fine sand 10 contained no characteristic flysch components that would suggest an input from Planinsko polje. On the other hand, quartz grains with abundant included organic matter prove an origin to the northwest, i.e. the Hotedrica region. THE SITUATION IN RAVNIK If only the accessible caves and general hydrology are considered, not much more can be said than is shown in Figure 1. With the exception of some insignificant, phreatic, fragments, there are no accessible non-vadose caves. The only indications of karst water streaming beneath Ravnik are highly active11 collapse dolines in various states of UTERIÈ, F. development. In some of them, such as Laka kukava and Dolga doline, there are strong blowing holes that at least testify to the existence of large caverns below. With one exception, the floors of all the collapse dolines are higher than the flood water level at Planinsko polje. Only Laka kukava reaches down to the level of 420 m12 a.s.l. On November 26, 2000, at the peak of one of the highest recorded floods at Planinsko polje, the lowest position in the doline was completely dry, without any trace of water between/below the boulders. So, it can be concluded that the water table was at an elevation of about 418m or lower. Logarèek was not visited on the same day, but according to the flood marks on the walls, the highest floods reached an elevation of about 447 to 449 m. Consequently, an estimation that there must have been a vertical water level difference of about a 30m is fair. Except for some isolated collapse dolines, most of them are arranged in lines. In 1954/55 MICHLER published a map, from which it is clear that the main stream line below Ravnik should simply be a joining of the shortest connections between the main collapse dolines. It must be noted that, decades before, many authors remarked that it is self evident that large collapse dolines in Ravnik indicate the underground stream of Ljubljanica. Michlers map was merely 10 Courtesy of Miran Nagode, Jamarski klub Logatec. 11 In the sense that the scree material is evidently creeping toward one or two foci at the bottom, and disappearing downwards. 12 The doline was surveyed by compass and metre tape in 1972 (UTERIÈ, 1973). The vertical difference of 106m between the highest elevation on the lip, and the lowest position at the bottom was obtained. This time, the maps at the scale 1:5000 were not available for public use, and the polygon was not related to the national grid system. According to the presently available map of 1:5000 the highest elevation is about 526m, and the bottom thus becomes 420m. RMZ-M&G 2002, 49 Where does Underground Ljubljanica Flow? a rather fanciful essay to elaborate common knowledge. Recently, UTERIÈ (1998; UTERIÈ ET. AL., 2001) demonstrated that the positions of the large collapse dolines indicate tectonic, rather than purely hydrogeological control. At first glance the Laka kukava string (Figure 2 / J) appears to be simply a continuation of the Slaven dol string, though rather disrupted at the approximate position that it crosses lineament13 M. However, closer inspection revealed that the two strings have little in common. The more than 100 unroofed caves in part of Ravnik, are the best studied within the Ljubljanica catchment area ( UTERIÈ , 1998), and their sedimentary fill is considered later in this paper. They all formed phreatically within the saturated zone and, as a rule, they lie along the contacts between dolomite and limestone beds within a sequence of thin interbedded layers (PEZDIÈ ET AL., o.c., Fig. 1). DISCUSSION OF ACCESSIBLE CAVE SYSTEMS Along the margins of Planinsko polje, the ponor caves are small, evidently young, and appear to be only a rudimentary adaptation to a very recent situation (UTERIÈ & PUC, 1970). The tectonically related Slaven dol string is the delimiter between two speleogically different areas (Figure 3). In the southern area (fed by the southern ponors), there are only 71 a few non-vadose caves. Among these only one short fragment of the top of a phreatic jump (Brezno nad Slavendolom, cad. no, 491), is purely phreatic, whereas others appear to be at least partly epiphreatic. Logarèek is the only fragment large enough to present some information about the overall system. Its organization clearly displays that it has passed through a phase of epiphreatic, perhaps even paragenetic, shaping. This would be possible only if the sinking river brought in a mechanical load, abraded the walls (and the ceiling), and filled the lower passages within the tier with gravel, before eventually forming a tunnel cave. If compared to Najdena, the lack of abandoned, dry, phreatic passages, as well as a near absence of fragments of such caves in the neighborhood, raises the question of whether any extensive phreatic maze had developed before the epiphreatic phase. The present state of knowledge implies no large-scale development of this type. Ignoring the previously mentioned exceptions, the situation is literally reversed in the northern area (fed by the northern ponors). Najdena jama is not the only phreatic development. All of the fragmentary caves that are not too deformed by slab spalling or collapse, as well as the many unroofed caves, display a clear picture of an unconstrained phreas. The question of how this contrast is possible within such a short distance is obvious. If such a pattern had developed only at the approximate level of the polje, there is a pos- 13 At the location, where the fault zone and the lineament come into contact, a broken zone about 100m wide was mapped along the latter, and evidently continuing in both directions along it. RMZ-M&G 2002, 49 72 UTERIÈ, F. Figure 3. Cave patterns on the northern (outflow) side of Planinsko polje RMZ-M&G 2002, 49 Where does Underground Ljubljanica Flow? sible answer. During the cold phases of the Pleistocene, when gravel production was much greater than today, the river was capable of transporting its load into the southern ponors, but not into the northern ones. So, epiphreatic channels might have been formed downstream from the southern ponors, while the channels more to the North remained basically phreatic. There are, however, some strong objections to this explanation: The existence of huge flowstone mass at the present entrance to Najdena jama testifies that there was very little change during the last fifth of the Pleistocene, and formation of the system cannot be attributed to the remainder; The vertical range of the phreatic system (at least 150m) pushes the time of formation of the northern system so far into the past that the available time within the Pleistocene becomes too short14 ; Logarèek has developed in two distinct levels, even disregarding what might have been denuded away, or buried below the sediments. Thus, its fundamental organization cannot be the result of just one episode; These considerations invoke the need for a more considered explanation. The clear-cut difference between the types of caves on opposite sides of the Slaven dol string, is an indicator that, during the formation of the system(s), the string worked as a delimiter between two completely distinct 73 hydrological zones. It appears that this at least reflects different input areas, and possibly also that different areas were active at different times. The ground plan of Najdena jama (Figure 3) shows that the system is just being dragged into contact with Planinsko polje, whereas it must have formed in completely different, but currently unknown, conditions. On the other hand, there are indications that Planinsko polje is quite young (UTERIÈ, 1996), certainly younger than the northern system. Unfortunately, in both systems, older cave wall scallops have been destroyed by condensation water or corrosion beneath a loamy coating, to such an extent that even speculation on the basis of their direction is impossible. For current purpose a simple statement that there are two basically different cave systems must suffice. Further research is in progress. UNROOFED CHANNELS IN RAVNIK AND THEIR SEDIMENTS Only the sedimentological features of denuded caves are considered in this paper, as they seem to be better indicators of cave system history than do morphological features (UTERIÈ, in press). All of the denuded caverns on the surface of Laki Ravnik were originally deep phreatic, oblique (reflecting the dip of the enclosing strata), and relatively small. They are filled with a variety of clastic sediments, and localized flowstone deposits also occur. For more detail, refer to UTERIÈ, 1998, 2000). 14 ZÖTL (1989, p.499) wrote The springs of the Ljubljanica River are absolute local base level of corrosion and erosion. The connection and cave systems are paleokarst, and this is true to entire polje landscape. Unfortunately, this revolutionary statement was not further supported and has remained more or less ignored by other students of the area. RMZ-M&G 2002, 49 74 Basal fill The most widespread fill material is reddish brown (2.5 YR 4/4, 5 YR 4/4) loam with a minor admixture of relatively large oolitic bauxite pebbles (derived from the Late Triassic - Carnian - beds) and coarse clasts of black chert. Preliminary X-ray diffraction analysis revealed the sediment to be mostly muscovite/illite, plus mixed-layer clay minerals of illite/montmorillonite type, chlorite plus mixed-layer clay minerals of chlorite/ montmorillonite type, calcium montmorillonite, and diaspore plus gibbsite, or just traces of bauxite minerals (MIIÈ, 2000). At first glance a potential sediment source area in the present Cerknièica river basin appears obvious, but similar outcrops of bauxite and chert do also appear at other sites that are not significantly more remote. Recently, bauxite pebbles (but not chert) have been found at elevations higher than the presently known primary outcrop in the Cerknièica river basin. Because no flowstone is found on the underlying cave walls, such sediments appear to have been deposited in cave systems that were completely water-filled. These deposits are nearly ubiquitous, and lie beneath most of the other preserved cave sediments, so the sediment is referred to as basal fill (UTERIÈ, 1998). Basal fill appears to be most common in Ravnik, but it is not completely absent in adjacent areas. To the southeast, such sediment can be traced back to the present Cerknièica stream. Northwestwards, it becomes nearly absent north of Logatec town. UTERIÈ, F. Laminated sandstone The sediment in question is composed of fine-grained quartz sand, cemented by calcite, that was later partly disintegrated again by ongoing surface chemical weathering. Some lamination is evident in unaffected sandstone. Visual inspection reveals no significant differences from a number of similar occurrences distributed all over the Julian Alps and Dinaric Alps of Slovenia (HABIÈ, 1992). Recently, about a dozen occurrences of similar laminated sandstone were found about 10km to the northwest, in a building site on the northern outskirts of Logatec. There it become clear that the earlier idea of a greater age for the laminated sandstone (than that of the basal fill) was correct. At some locations, structures that can best be attributed to thixotropic soft sediment creep within the caves were found, while on some others (neo) tectonic deformation is more or less obvious. Unfortunately, the exposures were obliterated by heavy building machinery before all the available information was collected. Laminated sandstone appears to be ubiquitous, as it can be observed all across the general area of Planinsko polje, regardless of the zones discussed above. Coarse-grained cave sediments Outcrops of conglomerates that were deposited in now-denuded caves are much less common. Clast size varies greatly, from coarse sand to large pebbles a few centimetres in diameter. The clasts are predominantly of Upper Triassic dolomite (originating in the Cerknièica river basin), with an admixture of local rock clasts, which makes the flow direction unambiguous. RMZ-M&G 2002, 49 Where does Underground Ljubljanica Flow? 75 Figure 4. Sediment associations in the unroofed caves in Ravnik. Many of the dolomite pebbles were hollowed out during vadose diagenesis. Small quantities of bauxite and chert also occur, but it is not yet possible to say whether they were introduced as part of the stream load or simply admixed on the spot. Coarse rubble, originating from the channel roof or walls, is also present locally. Generally the conglomerate displays typically alluvial features, such as graded bedding and layered sedimentation. At some localities flowstone occurs with the conglomerate, and it appears that there were several alternating phases of conglomerate sedimentaRMZ-M&G 2002, 49 tion and flowstone deposition. As flowstone is an indicator of subaerial conditions, its mere existence indicates that conditions in the cave system were no longer deep phreatic at the time of conglomerate sedimentation. In Ravnik, conglomerate-filled caverns are restricted to the area between lineaments M and N. Half a dozen kilometres northwest, at Logatec, on the northwestern side of lineament K, the conglomerate area widens. In parallel, pebbles of Permian sandstone, originating from the northwest, become more and more commonly admixed. As the prevailing component, arriving from the same 76 direction, comprise late Triassic dolomite pebbles, appearing exactly the same as those originating to the southeast, in the Cerknièica river basin, the location of mixing and the relative proportions involved are difficult to establish. Flowstone In Ravnik, conglomerate and flowstone were deposited more or less simultaneously, into channels previously washed partly clean of the basal fill. Relatively dry periods of flowstone deposition alternated with possibly catastrophic flood events, during which gravel-laden water rushed into dry caves. The length of the time gap between basal fill emplacement and the last phase of partial outwash and deposition is uncertain. Determination of the age of the flowstone deposits in the denuded caverns is clearly beyond the reach of the U-Th series method. Following a single failed attempt, at Bergen University (MIHEVC, 2000), no further direct attempts have been made to measure their date(s) of deposition. The dolomite and bauxite pebbles, as well as the chert gravel within the conglomerate, do not differ from those in late Pleistocene and Recent Cerknièica river alluvial fan gravels in the Cerkniko polje. Thus, the cessation of conglomerate sedimentation in Ravnik is clearly a result of the inrush of the present Cerknièica into the Cerkniko polje basin, as it abandoned its old underground routes towards the northwest15 . The oldest UTERIÈ, F. directly dated sediment within the present Cerknièica alluvial fan is 55ka old (GOSPODARIÈ AND HABIÈ, 1979, p.43), whereas indirect observations push the beginning of the sedimentation back to about 80ka b.p. On the other hand, conglomerate-filled caves are exposed in the slopes of most of the larger collapse dolines, which suffered their essential transformation during the last glacial episode (UTERIÈ, 1998). So, it may be concluded that conglomerate (and flowstone) sedimentation took place during the Pleistocene, but not at its very end. ORGANIZATION OF THE DRAINAGE SYSTEM From consideration of the previous text one conclusion can very clearly be reached. Some tectonic structures (including lineaments) act as strong delimiters between areas (blocks) of completely different speleological and hydrological properties. a16 . Area northeast from lineament N and southeast from lineament L The basal fill is preserved at high elevations, whereas laminated sandstone appears at lower ones, and no conglomerate has yet been found in this area. In areas where laminated sandstone has not been found, there are some larger caverns, which are accessible to humans. Close to the caverns are larger, unroofed, caves containing basal fill (GERL ET. AL., 1999). Absence of conglomerate and the presence of accessible caverns are charac- 15 One of the consequences of this catastrophic event is retransportation of chert from Planinska jama to Najdena jama (UTERIÈ ET. AL.., in press). 16 See Figure 2 RMZ-M&G 2002, 49 Where does Underground Ljubljanica Flow? teristic. If any collapse dolines exist in this area, they were excluded from the drainage system so long ago that they have completely lost their characteristic geometry. Thus, remnants of old underground karstification exist, but there are no underground indications of active karst streams. b. Area between lineaments M and N, and southeast from lineament K This area is characterized by a number of unroofed, originally phreatic caves, predominantly associated with alternating limestone/ dolomite beds of the Malmian (UTERIÈ, 1998). Though laminated sandstone is not completely absent, it appears to have had been mostly washed out and replaced with basal fill. Much later a minor part of the basal fill was removed and conglomerate plus flowstone took its place. The direction of input of the basal fill is quite likely from the southeast, but direct proof has not yet been found. The conglomerate, however, contains a minor admixture of local rocks, and it is easy to trace a source to the southeast, in the present Cerknièica river basin. Highly active collapse dolines, like Vodika dolina, Rakovska Kukava and Laka Kukava, testify that active underground streams (or perhaps a single stream) exist beneath Ravnik. Water tracing experiments do not permit any other interpretation than the present southeast-northwest flow direction. The great potential difference (Dh ≥ 28 m) across lineament M, between Logarèek and Laka kukava, indicates that the system is highly transmissive. So, it may be presumed that the present flow corridor makes use of older passages, most likely the ones that were previously filled with basal fill, and later washed out. RMZ-M&G 2002, 49 77 As they are bound intimately to the Malmian, it is possible simply to extrapolate this sequence below the elevation of 420m a.s.l., and east of lineament M. Since the bedding is oblique to the general direction of Ravnik, such a procedure is possible only in the area where the Malmian is trapped between the two lineaments, which fortunately is in the main area of interest. Before positioning the flow corridor positively, two constraints must be considered. Firstly, the block between the two lineaments has been broken down into slices, one or two hundred metres wide, and affected by sinistral displacement of several dozen metres (UTERIÈ , 1997). Consequently, the overall, effective dip of the rock does not equal the one measured by compass. Secondly, there is an optimum depth for cave channel formation (WORTHINGTON, 2001). So, very detailed mapping in the critical areas might provide data, sufficient for relatively exact determination of the main channels location/depth, below the present water table. The role of the apparent continuation of the Slaven dol string (L) remains rather unclear. Closer to Planinsko polje, southwest of lineament M, it is a delimiter between two completely different cave systems. North-east from it, however, such a difference has not been observed, at least not as regards unroofed caves. Another question is its influence on the water table. The previously mentioned absence of flood water at the base of the doline could mean either that the general level of the water table on both sides of the doline is this low, or that it is this low only on the downstream side. The volume of the doline (4.17Mm3, UTERIÈ, 2000-b, 224) suggests that water has been penetrating through the broken zone for a long time. 78 Thus, collapsing possibly has no significant influence on the actual transmissivity, and the water table might be low on both sides of the doline. cs. Area southwest of lineament M, and southeast of lineament Slavendol string (L) The area in question covers the southern ponors of Planinsko polje and adjacent areas downstream. Hard information about the water body is available only from the Logarèek cave. During flood periods the lower parts of the cave are water filled to a level close to that of the water table on the polje, and high above the still dry floor of Laka kukava. During summer drought the tested ponors feeding Logarèek (see footnote no.7!) become dry. As described above, at this time colder, winter, water remains trapped in the areas closer to the polje, whereas the more northern parts are influenced by warm water arriving more from the south. In general, passages of Logarèek run parallel to the Slaven dol string on its eastern side, only a few hundred metres away (UTERIÈ ET. AL, 2001, 15, Figure 3). So it may be concluded that the main flow corridor funnels into the angle between lineament M and the Slaven dol string (L). In this area there are some active collapse dolines, one of them exactly at the intersection of the two tectonic lines. This makes the option that the stream corridor crosses lineament M seem more plausible than the alternative possibility that it runs northwest, parallel to the lineament. cn. Area southwest of lineament M and northwest of the Slavendol string (L) This sector is generally believed to transmit the greatest quantities of floodwater (see UTERIÈ, F. above). However, Brezniks (o.c.) estimation is not the only one, and figures from other authors vary by a factor larger than two (PUTICK, 1888; JENKO, 1954 [cit. GAMS, 1963]; HABIÈ, 1976). It must be noted that none of the swallow hole capacities has been obtained by direct measurement at their openings, but only by calculating the riverbed water budgets. As the water level must increase significantly before the northern ponors receive water at all, BREZNIKS (o.c.) statement would be better as: When the water level is so high that all the northern ponors become really active, the polje margins can absorb 40m3s-1 more than they do when the water level is lower. Estimations made by various authors thus imply that, as might be expected, the capacities of the ponors change according to the hydrological conditions in the polje. It must also be noted that most of the accessible cave passages (UTERIÈ, 1994), as well as unroofed caves, lie along the upper or lower margins of two massive dolomite packages (Figure 3). So, it is perhaps better to suggest two transmission systems within the northern zone. Some information about the water body can be obtained within the western area, where most of Najdena jama, Vranja jama and Gradinica are situated. As demonstrated above, it is not likely that great quantities of water would flow through there. The origin of the sandy sediment in Gradinica indicates that the cave is displaced from, i.e. west of, the main flow corridor of the waters sinking in Planinsko polje. Consequently, the floodwater observed there could simply be dammed by a main stream farther to the east. A northward-trending string of large, active RMZ-M&G 2002, 49 Where does Underground Ljubljanica Flow? collapse dolines (= Smreènica string) lies to the northwest of Najdena jama, implying that the flow corridor must cross a tectonic line. The northerly orientation of the string would permit both flow directions, either west to east, or the opposite. Weak indications in the northernmost part of Najdena jama imply that the latter option is more likely. On the other hand, if this were true, the stream should turn north very soon after crossing the string. If not, it would unavoidably influence sedimentation in Gradinica. Without more data there is no point hypothesizing further. Relative rapid reactions of the water level in Gradinica to the flooding/emptying of the polje indicate that an obstacle, not important for small discharges, but with limited transmissivity for larger discharges, must exist close by. So, as in block cs, it might be expected that the main stream runs quite close to lineament M, parallel to it on its southwestern side, until it finds a weak point, and the water pours into area d. So, it would be expected that most of the water leaves the polje through channels related to the eastern dolomite pack. Unfortunately, the water table is not accessible there, and active collapse dolines appear to be absent. d. Area northeast from lineament M, and northwest from lineament K This area has not been researched in great detail, nor have the unroofed caves been studied systematically. Being situated so far to the northwest, it is the most probable location for the main flow corridors to turn northeastwards, towards the Ljubljanica springs at Vrhnika. Two aspects must be underlined. RMZ-M&G 2002, 49 79 In this area sedimentation within the unroofed caves becomes influenced more and more by input from nonkarstic area farther to the north and northwest. Only at first glance is the transition gradual. In fact, channels filled with various materials of very different origin, interfinger in 3-D. This implies an intensive interchange of sedimentation, washing and resedimentation from both sides, but with intensive oscillations in the water regime too. In an area of less than 2km2, in the northern angle of the crossing of lineaments M and K, there are about two dozen mostly very active collapse dolines with the general name Logake Kolievke. They differ from those previously mentioned in that they are not arranged only along strings, but also into a planar, clustered pattern. Such organization is not very common. Undoubtedly it indicates intensive fracturing of the parent rock (PEÈERKIN & KATAEV, 1985) as well as strong water streaming below. The conclusion that they impede the drainage of block cn appears to be obvious. Block b (and indirectly, block cs) appears to be better drained, and it is possible that their main outlet turns northeastwards before crossing lineament K. CONCLUSIONS The former discussion can be summed up in the following statements: A relatively direct and only slightly impeded flow corridor runs between lineaments M and N, approximately along the axis of Ravnik, perhaps corresponding with links A1 and B on Figure 1. Though the Rakovska kukava and Laka kukava collapse dolines indicate 80 UTERIÈ, F. strong, concentrated flow, observations of the unroofed caves give an impression that a well-developed maze exists, so that collapses cannot influence the bulk transmissivity of the flow corridor. The southern ponors group drains the polje quite directly, but after encountering lineament M, water turns northwestwards along it. At the intersection of lineament M and the Slavendol dol string at least a part of the flow breaks through lineament Ms broken zone and pours into the formerly mentioned flow corridor. The northern ponors group seems not to be very uniform, and nor can it take as great a proportion of the flood water as was formerly believed. The system was formed in hydrological conditions that might differ from the present ones, to such an extent that it hasnt yet adapted completely. The main flow corridor runs parallel to lineament M until close to Gradinica, and then turns northeast, having crossed the continuation of lineament K. Gradinica lies in a dead corner of the system, and it does not transmit large if any at all - quantities of water originating from Planinsko polje. Regional lineaments, with 100m-wide, or wider, fracture zones generally serve as delimiters between blocks with distinctive, local, karst hydrology. Sinistral strike-slip faults, with fracture zones not wider than 20m do not influence underground streaming to any great extent. On the other hand, they are the most prone for the location of extremely large dolines, opened up by ongoing undercutting, rather than by a sudden collapse. In this configuration, the Slaven dol string appears to be an exception, as it displays properties of both of the previously listed structural lines. Further research is in progress. The locations of large collapse dolines are primarily important as indicators of weakened rock, and only secondarily as indicators of the flow corridor(s) below. POVZETEK Kje teèe podzemska Ljubljanica? Kraka Ljubljanica je e vsaj 150 let v sredièu zanimanja krasoslovcev, najbolj pa pozornost privlaèijo delno e neznane poti, po katerih se ponikalnica pretaka med posameznimi krakimi polji. e posebej izstopa odsek med Planinskim poljem in Vrhnikimi izviri, kjer je znanih sorazmerno malo vodnih jam. Kljub temu, da v zadnjnih letih ni prilo do veèjih posmeznih odkritij, lahko na osnovi izboljanega splonega poznavanja ustroja in delovanja krakih pretoènih sistemov ter posameznih na videz nepovezanih terenskih opazovanj pridemo do uporabnih zakljuèkov. Pri tem izhajamo iz spoznanja, da se podzemske vode pretakajo, oz. oblikujejo kanale (jame) vzdol jasno definiranih pretoènih koridorjev, ki jih doloèajo veèji prelomi (oz. lineamenti, èe jih ugotavljamo samo posredno) in litologija. Pri slednjem izhajamo iz spoznanja, da je na obravnavanem obmoèju v znanih podzemskih in brezstropih (denudiranih) jamah veèina kanalov vezanih na stik apnenec-dolomit. Sedimenti v brezstropih jamah nudijo informacijo o smereh pretakanja oz. izvornih obmoèjih v preteklost, s tem pa posredno pomagajo RMZ-M&G 2002, 49 Where does Underground Ljubljanica Flow? ovrednotiti hidrogeoloko vlogo posameznih lineamentov. Na ozemlju med Planinskim poljem in Ljubljanskim Barjem najbolj izstopata lineamenta M in N, ki v dinarski smeri, vzporedno z Idrijskim prelomom, potekata vzdol Begunjsko-Logakega Ravnika. Tako so zaèrtani trije pretoèni koridorji, ki topoloko ustrezajo rezultatom sledilnih poskusov v preteklosti. Povprek na te strukture leita lineamenta K in L, ki v preènodinarski smeri razmejujeta posmezne hidrogeoloko individualne bloke. Lineament L loèi enoti, ki izkazujeta popolnoma razlièno spleogenezo v preteklosti, danes pa kjer sta v neposrednem stiku s Planinskim poljem - precej razliène poiralne lastnosti. Jugovzhodneji blok, ki je sposoben odvesti nizke in veèino srednjih voda Unice, je v preteklosti doivel fazo epifreatiènega oblikovanja, kar z drugimi besedami pomeni, da je ponornica vnaala mehansko plavje s povrja. Obratno pa je krako podzemlje v severozahodnem bloku popolnoma freatièno. Dosedanja literura meni, da so najpomembneji ponori v tem bloku, medtem ko avtor ugotavlja, da je to verjetno le videz. Zdi se, da se krako podzemlje v tem bloku ni razvijalo v povezavi s Planinskim poljem in da se v njegovo hidrologijo vkljuèuje ele v najmlajem èasu. Ugotovitve povzamemo v sledeèe: Sorazmerno neposreden in le nekoliko oviran pretoèni koridor poteka med lineamentoma M in N. Èeprav posmezne udornice kaejo na krajevno koncentriran tok, je tu bolj verjeten labirinten splet krakih kanalov, kjer pretoka vode posamezni udori stropov bistveno ne moteijo. RMZ-M&G 2002, 49 81 Juna skupina ponorov odvodnjava Planinsko polje precej neposredno. Ko pa tok trèi na lineament M, zavije vzdol njega proti sevrozahodu. Na seèièu z lineamentom L se vsaj del vode prebije skozi lineament M in vkljuèi v prej omenjeni pretoèni koridor. Severna skupina ponorov ni niti enotna, niti tako uèinkovita, kot se zdi. Jamski splet je nastal v razmerah, ki so od dananjih verjetno precej razlikovale in se slednjim e ni popolnoma prilagodil. Glavni pretoèni koridor poteka vzporedno z lineamentom M do bliine jame Gradinica, kjer zavije proti severovzhodu, èim preèka lineament K. Gradinica lei v mrtvem kotu sistema in ne prevaja veèjih kolièin voda, poniklih na Planinskem polju. Regionalni lineamenti, s po100m in veè irokimi zdrobljenimi conami, so veèinoma izrazite loènice med bloki, ki kaejo samosvojo krako hidrologijo. Levozmièni prelomi, katerih zdrobljene cone niso ire od 20m, bistveno ne vplivajo na podzemsko pretakanje. Po drugi plati so ugodni za nastanek veèjih udornic, ki pa ne nastajajo s hipnim zruenjem stropa, temveè poèasnim spodkopavanjem. V tem kontekstu je lineament L izjema. Podrobno preuèevanje je v teku. Mesta nastanka velikih udornic v prvi vrsti odraajo oslabljeno kamnino, manj neposredno pa so udornice pomembne kot kazalci pretoènih koridorjev v globini. 82 UTERIÈ, F. REFERENCES BRENÈIÈ, M. (1992): Koelevc (Summary17 ). Nae jame 34, 41 - 51. BREZNIK, M. (1998): Storage reservoirs and deep wells in karst. Balkema, 1-251, Rotterdam. CENTRAL CAVE REGISTER OF SLOVENIA, maintained by the Speleological Association of Slovenia and the Karst Research Institute ZRC SAZU. ÈAR, J. (1982): Geologic setting of the Planinsko polje ponor area (Summary). Acta carsologica, 10 (1981), 75-105. GERL., M., STEPINIK, U., UTERIÈ, S. (1999): The unroofed cave near the bunker (Laki Ravnik) (Summary). Acta carsologica, 28, 77-90. FORD, D. C., WILLIAMS, P.W. (1989): Karst geomorphology and hydrology. Unwin Hyman, 1 - 601, London. GAMS, I. (1959): About the situation and the genesis of the longest slovene caves (Summary). Nae jame 1 (1), 4-10. GAMS, I. (1963): Logarèek (Summary). Acta carsologica, 3, 1-84. GAMS, I. (1966): Factors and dynamics of corrosion of the carbonatic rocks in the Dinaric and Alpine karst of Slovenia (Summary). Geografski vestnik, 38, 11-68. GOSPODARIÈ, R. (1976): The Qaurternary caves development between the Pivka basin and Polje of Planina (Summary). Acta carsologica, 7, 8-135. GOSPODARIÈ, R. (1982-a): Stratigraphy of cave sediments in the Najdena jama at Planinsko polje. Acta carsologica, 10, 173-193. GOSPODARIÈ, R. (1982-b): Morphologic and geologic situation of caves at the Planina polje ponor boundary. Acta carsologica, 10, 157-171. GOSPODARIÈ, R., AND HABIÈ, P. (1976): Underground water tracing. Institute for the Karst Research SAZU, Postojna, 1-309, Postojna. GOSPODARIÈ, R., AND HABIÈ, P. (1979): Karst phenomena of Cerkniko polje (Summary). Acta carsologica, 8, 7-162. HABIÈ, P. (1963): Dolines en forme de puits, dites kolievke, et le cours deau soutterain (Résumé). Treæi jugoslavenski speloloki kongres. Speleoloki savez Jugoslavije, Sarajevo, 1-272. HABIÈ, P. (1992): Les phénomenes paléokarstiques du karst alpin et dinarique en Slovénie. In: Karst et évolutions climatiques: hommage à Jean Nicod, Salomon, J.-N. and Maire, R., eds., Presses Universitaires de Bordeaux, Bordeaux, 412-429. 17 The titles of summaries/abstracts (if they exist) are given just to show the foreign reader the contents of the original texts, which are, however, considered in the whole. RMZ-M&G 2002, 49 Where does Underground Ljubljanica Flow? 83 JENKO, F. (1976): Hidrogeologija poreèja Ljubljanice. Vodnogospodarska osnova poreèja Ljubljanice. Prirodne osnove. Hidrogeologija. Projekt nizke zgradbe, Ljubljana, [Cit: I. Gams, 1963]. KRIVIC, P., VERBOVEK, R., DROBNE, F. (1976): Hidrogeoloka karta. Geoloki zavod, Ljubljana. LAJOVIC, A. (2001): Personal communication. LOWE, D. J., GUNN, J. (1997): Carbonate speleogensis: an inception horizon hypothesis, Acta carsologica, 26, 457-488. MARUSSIG, M., VELKOVRH, F. (1959): Gradinica (Zusammenfassung). Nae jame, 1 (1), 24 - 28. MICHLER, I. (1954-55): Planinska dolina. Proteus, 17, 10-1. MIHEVC, A. (1979): Geomorfoloka karta ozemlja Logakih Rovt, diplomsko delo. FF UL, 1-63. MIHEVC, A. (1976): The cave Brezstropa jama near Povir (Abstract) Nae jame, 38, 65-75. MIHEVC, A. (2000): Personal communication. MIIÈ, M. (2000): Personal communication. PEZDIÈ, J., UTERIÈ, F., MIIÈ, M. (1998): On the role of clay-carbonate reactions in speleo-inception, Acta carsologica, 17 (1), 187-200. PALÈIÈ, T. (2001): E-mail message, 15 Jan 2001. PEÈERKIN, A. I., KATAEV, V. N. (1985): O vzjamosvjazi raspredelenija karstovyh provalov Palankaresa i Kanjada delj Ojo (Ispanija) s treèinovatostju i napraennym sostajeniem massiva. Geologija i razvedka, (1), 55-60. PUTICK, W. (1887): Die unterirdischen Flussläufe von Inner-Krain Das Flussgebiet der Laibach. Mitt. KuK Geogr. Ges. Wien, 33, 5-6, 277-289. PUTICK, W. (1889): Die Die Katavotrons im Kesselthale von Planina in Krain. Separat-Abdruck von Vochenschrift des österreichischen Ingenieur und Architekten Vereines, 46-47, 1-7. UTERIÈ, F. (1973): On the problems of collapse dolinas and allied forms of high Notranjsko (Southcentral Slovenia) (Summary). Geografski vestnik, 45, 71-86. UTERIÈ, F. (1976): Kvartarni sedimenti v zasutih breznih Notranjskega krasa, diplomsko delo. FNT UL, 1-91. UTERIÈ, F. (1982): Mophology and hydrography of Najdena jama (Summary). Acta carsologica, 10 (1981), 127-156. UTERIÈ, F. (1991): On the objective of speleological science (Summary), Nae jame, 33, 73-85. UTERIÈ, F. (1994): The Kloka cave and speleo-inception (Summary). Nae jame, 36, 9-30. UTERIÈ, F. (1997): Poljes and caves of Notranjska, Acta carsologica, 25: 251-289 (1997). RMZ-M&G 2002, 49 84 UTERIÈ, F. UTERIÈ, F. (1997): Rakovska kukava - collapse or tumour doline? Acta carsologica, 25, 251-289. UTERIÈ, F. (1998): Interaction between the cave system and the lowering karst surface. Case study: Laki Ravnik. Acta carsologica, 27 (3), 115-138. UTERIÈ, F. (1999): Vertical zonation of the speleogenetic space, Acta carsologica, 28, 187-201. UTERIÈ, F. (2000-a): Speleogenesis in the Ljubljanica river drainage basin, Slovenia. In A.B. Klimchouk, D.C.Ford, A.N. Palmer, W. Dreybrodt (Eds.): Spelogenesis, Evolution of Karst Aquifers, National speleological society, 397-406, Huntsville. UTERIÈ, F. (2000-b): Are collpase dolines formed only by collapse?, Acta carsologica, 29 (2), 213230. UTERIÈ, F., in press: Cave sediments and denuded caverns, in: Sasowsky, I.D. and Mylroie, J. (eds.), Studies of Cave Sediments: Kluwer Academic/Plenum, New York (page numbers not yet affixed). UTERIÈ, F., PUC, M. (1970): The karst underground by the NE nook of the Planinsko polje (Summary). Acta carsologica, 5, 205-270. UTERIÈ, F., JAKOPIN, P. (1979): Kvantitativno preuèevanje elementov fiziène speleologije v prostoru Planinskega polja. Elaborat v arhivu IZRK ZRC SAZU, Postojna. UTERIÈ, F., ÈAR, J., EBELA, S. (2001): Collector channels and deflector faults (Summary). Nae jame, 43, 8-22. WORTHINGTON, S. R. H. (2001): Depth of conduit flow in unconfined carbonate aquifers. Geology, 29 (4), 335-338. ZÖTL, J. (1989): Paleokarst as an important hydrogeological factor. In: BOSAK, P., FORD, D.C., GLAZEK, J., HORAÈEK, I., (eds.): Paleokarst, Academia, 483509, Prague. RMZ-M&G 2002, 49