JANKOVICHIAN, SZELETIAN OR A LEAF POINT INDUSTRY: ANALYSIS OF THREE SMALL LITHIC ASSEMBLAGES

: The eponymous site of the Jankovichian industry was found at the Öreg kő cliff, in the northern part of the Transdanubia, Western Hungary. From the thick layer complex of the Jankovich cave, however, only 104 lithics were collected and the scarce data showed that the pieces belong to several archaeological entities. At the same time, the nearly total lack of the field documentation allowed the reconstruction of the stratigraphic position of each artefact only in a few cases. The stratigraphic integrity is missing from the archaeological material of the Kiskevély and Szelim caves as well as the Csákvár rock shelter, and according to the recent evaluations the bifacial tools from the Dzeravá skála (Pálffy cave) and Lovas belong to the Micoquian and the Late Pa­ laeolithic period.In this paper we analyse the following three assemblages, excavated after World War II in Transdanubia: the Pilisszántó rock shelter II, the Bivak and the Remete Felső caves. The chronological, basically, bio­stratigraphic data known from these layers are also evaluated. The conclusion of the study is that (1) the chronological data of the studied sites do not permit to place the archaeological occupation of each cave into the Early Würm or to the Late Middle Palaeolithic period and (2) the validity of a distinct Jankovichian industry cannot be proved.

Even in Hungary, a number of different assemblages and finds are united under this term, including the industries of the Szeleta cave, 7 the Szeletian industry as it was described by F. Prošek in 1953 8 and the leaf shaped artefacts selected from the mixed surface collections. 9In the 1950s 10 two geographical groups of the Szeleta culture were differentiated in the Bükk Mountains and Transdanubia, respectively.As a result of the excavations in the Remete Upper cave (1969-1971) instead of this later Transdanubian group the Middle Palaeolithic Jankovichian industry was postulated. 11The definition of this new cultural entity, dated to the Early Würm ('Altwürmʼ, MIS 5 da) by archaeological, paleontological and palaeobotanical arguments was based on 176 lithic artefacts, collected from eight different localities in Hungary and one in Slovakia.However, the richest and eponymous site yielded only 104 lithics, most probably belonging to several different archaeological industries, excavated from the 6 m thick layer sequence and only at five pieces are the data on the exact place of the recovery available. 12The field observations are also absent from the Szelim 13 and Kiskevély 14 caves, excavated before the World War II.Similarly the history of the interpretation of the Csákvár rock shelter and the Lovas ochre mine locality are typical and instruc tive.During the palaeontological excavations of the Csákvár rock shelter in 1926 and 1928 a pierced deer canine and a human metacarpal bone was found in the Pleistocene light brown loam. 15In 1951 further artefacts and another human bone were collected from the backdirt of the previous field works. 16The age of the rather atypical fauna was estimated to be more recent than the characteristic 'Würm I' assemblages but older than the 'Würm III' faunas (MIS 4 and MIS 2, respectively). 17mong the lithics collected in 1951 L. Vértes compared a bifacially worked knife and a side scraper to the Mousterian tool types from Tata and he noted that a single flake with large bulb of percussion is similar to the blanks known from the Jankovich cave. 18This atypical piece was enumerated among the Jankovichian finds in the eight ies 19 and in 1993 the artefacts identified earlier as Mousterian tools by GáboriCsánk as well as two flakes were also listed among the lithics of the same industry. 20In our view, however, this new classification of the scattered finds from the Csákvár rock shelter is not convincing enough.
Similarly, the single leaf shaped scraper of limnic quartzite 21 excavated at Lovas allowed to describe this locality as the only one openair site of the Transdanubian Szeletian.Although in the 1970s and 1980s M. Gábori and V. GáboriCsánk suggested, that the presence of a single typical tool was not sufficient to list the site among the Jankovichtype industry, 22 in the monograph consecrated to this entity twelve flakes, a core fragment and a raw material fragment, each made of radiolarite were also enumerated among the Middle Palaeolithic artefacts 23 and the single typical tool was compared to the earliest pieces of the industry known from the Kiskevély cave. 24ecently M. PatouMathis summarizing her observations on the osseous artefacts suggested that this local ity has been repeatedly used from the Middle Palaeolithic until the recent Prehistoric times, 25 and pointed out that the artefacts were collected from three features probably representing different periods.However, the data given by Gy.Mészáros and L. Vértes clearly show that the leaf shaped scraper was found in layer 5 of feature 2, together with the radiolarite flakes and the doublebevelled point, 26 as well as several bone implements made of elk ulna, 27 pseudometapodial awls, 28 and finally, the single tool made of ibex bone. 29he Lovas site was originally dated to the Würm I-II interstadial, later to the Early Würm (Varbó phase, following the mammal biostratigraphy developed in Hungary, MIS 5a) 30 or to the Middle or Upper Würm (Istállóskő phase). 31In our view, the leaf shaped scraper was most probably found in a secondary position and does not belong to the otherwise homogenous Late Palaeolithic assemblage of the characteristic, specialised bone tools associated with reliable radiocarbon dates. 32Anyway, systematic taphonomic, typological and technological studies of this important assemblage will be necessary in the future.
Finally, from the Pálffy/Dzeravá skála cave (Western Carpathians, Slovakia) only the single bifacially worked piece excavated by J. Hillebrand was originally mentioned by V. GáboriCsánk. 33Later the presence of the Levallois flaking, 34 leaf shaped scrapers and pieces similar to the Faustkeilblatttype were also identified on the drawings published by F. Prošek.At the same time, some artefacts were compared to the tools known from the Kiskevély 35 and the Jankovich cave.Consequently, GáboriCsánk classified each 'Szeletian' lithic from this cave as belonging to the Jankovichian industry. 36Recently, however, the presence of both the Szeletian or Jankovichian artefacts in the Dzeravá skála assemblage were questioned and the single bifacially worked tool found during the 2002-2003 excavations in the middle part of layer 11 was classified as belonging to the Micoquian industry after an infinite radiocarbon date from the same level. 37n the last ten years the problem of the Jankovichtype artefacts was discussed by Zs.Mester in connection with the socalled technological investigations of the bifacial tools from the Szeleta and the Jankovich caves. 38The result of the investigations suggested that the asymmetric forms, generally made on flakes, and considered as the characteristic forms of the lower layer of the Szeleta cave are identical with pieces of the Jankovich cave.Earlier 39 we pointed out several problems concerning the low number of studied pieces in each assemblage, 40 the pointless use of percentages when the studied population is well below 100 elements and the contradictions at the strati graphic interpretation of certain layers, especially at layer 4 of the Szeleta cave. 4134 The use of this method was observed during the recent evaluation of the assemblage too: kaMinská et al. 2005, 41, 45, Fig. 25,5. 35 The given tools from the Kiskevély cave were classified earlier as belonging to the Mousterian industry (see note 14) and they are very similar, however, to the Tata artefacts.
36 Gábori-Csánk 1993, 80. -In the same volume the bifa cially worked tool excavated by J. Hillebrand in the Pálffy cave/Dze ravá skála was listed among the tools from the Jankovich cave: Gábori-Csánk 1993, Pl. Ia-b. 3. -cf. Markó 2013a, 17, note 29. 37 kaMinská et al. 2005, 55;kaMinská 2014, 94-95. 38 Mester 2010; Mester 2011; Mester 2014a. 39Markó 2016. 40For instance, in a recent paper Zs.Mester claimed that seven of the 18 asymmetric pieces excavated in the Szeleta cave were found in the upper layers 5, 6 and 6a, nine artefacts in the lower layers 2, 3 and 4, and finally, no stratigraphic data are available in the case of further seven pieces.This way, however, the number of the asym metric pieces is not 18 but 23. and in fact, he analysed 17 leaf shaped implements: Mester 2017, 78.-It is embarrassing, that in a paper published in the preceding year the number of the asymmetrical pieces was 13, seven of them were from layers 5 and 6 and six pieces from layers 4 and 3: lenGyel et al. 2016, Table 5. 41 Mester dated this assemblage to the early Szeletian.This is, however, inconsistent with the data presented in the same papers, as the majority of the leaf shaped artefacts from this stratigraphic unit belong to the group characteristic for the Evolved Szeletian: Mester 2011, tabl.4; Mester 2014a, Tabl. 4. -c.f.Markó 2016, note 1.
In our view, the information available from the more than 100 years old excavations of the Szeleta is not sufficient to make well based conclusions in the 21st century. 42Furthermore, our studies shed some light to the role and intensity of the postdepositional effects, which could change basically the stratigraphic position of the given artefacts. 43Similarly, the Late Gravettian classification of the assemblage from layer 3 and 4 was rejected on the strength of the presence of the asymmetric leaf shaped points and the occurrences of the Gravettian tools were in terpreted as an admixture. 44At least in some cases, however, the error occurred after the excavations with the incor rect stratigraphic and typological determination of the artefacts. 45E STUDIED ASSEMBLAGES Ultimately, of the Transdanubian Seletian/Jankovichian sites very few or no stratigraphic information is available from the Jankovich, Kiskevély, Szelim and Csákvár caves, while the recent evaluation led to the conclu sion that the openair site of Lovas and the Pálffy/Dzeravá skála cave belong to different, Late Palaeolithic and Micoquian industry.These observations raised the question of the validity of the term 'Jankovichian'.In the present paper we analyse the available information from three localities excavated after World War II in the northeastern part of the Transdanubia: the Pilisszántó rock shelter II, the Bivak and the Remete Felső caves (Fig. 1), that may serve further evidences to this problem.

Pilisszántó Rock shelter II
The site is lying in the southern part of the Pilis hill, close to the bottom of the valley, at 386 m a.s.l.The infilling of the little chamber was partly excavated in 1946 by L. Vértes.In the Pleistocene layers only two atypical chipped stone artefacts of greenish radiolarite (Fig. 2.1) and Slovakian obsidian (Fig. 2.2) were found.The pieces were first dated to the Magdalenian, 46 later to the Transdanubian Szeletian 47 and finally to the Jankovichian 48 indus try.Recently the occurrence of obsidian as an extralocal raw material in this assemblage led K. Biró to date the site to Early Upper Palaeolithic. 49e implements were classified as Szeleta scrapers, 50 later scrapers with large bulb of percussion, similar to the Jankovich tools 51 and, rather surprisingly, as a unifacial leaf shaped scraper and a scraper. 52In our view, both pieces belong to the group of the heavily fragmented blanks partly showing traces of intentional modification, but partly shaped by natural factors.These forms, sometimes called as 'raclettes', are known not only from the Janko vich cave but also from the Dzeravá skála 53 and Szeleta 54 caves as well as from the lower ('Aurignacian I') layer of the Istállóskő cave in the Bükk Mountains, 55 clearly illustrating the problems with the typological classification of these partly naturally fragmented lithics.
During the excavations of the Pilisszántó rock shelter II several Pleistocene layers and traces of important erosional events were documented.According to the original report the artefacts were found in the lower layer group (layers 8-10). 56Later the lowermost, brownishred layer 10 57 or the lower brown layer 9 58 was given as the place of recovery of the artefacts.The recent stratigraphic evaluation of the site placed the occurrence of the lithics to the border of the reddish brown and the redbrown loam.-Importantly, the arte fact from the Jankovich cave, mentioned as similar to the piece made of obsidian (Fig. 2.2 in the present paper) is most probably the frag ment of a typical blade, collected from the Gravettian layer: Gábori-Csánk 1993, IX,9, VIII,19;cf. Markó 2013, Fig. 2, 3. 53 The pieces excavated by F. Prošek were enumerated among the artefacts from the uppermost level of layer 11: kaMinská  et al. 2005, 45-50, Fig. 28.  5Markó 2016, 16, 19, Table 3. 55 Markó 2015, 22.
57 Vértes 1955a, 270 -According to the excavation diary, this stratigraphic unit (Layer 9, following the revised layer sequence) was completely sterile in archaeological and paleontological point of view, see: dobosi-Vörös 1986, 27.
59 Layers 9 and 10 by Vértes or Layers 8 and 7 of the re vised sequence.Following the archaeological considerations, i.e. the presence of the Jankovichian industry it was suggested that the scrap ers were found in the uppermost horizon of the reddish brown Layer 8: dobosi-Vörös 1986, 30, Fig. 1.
The formation of the reddish brown layer was estimated to the Würm I or PreWürm, however, as in this layer only a single indifferent bone fragment was found this age was based exclusively on the presence of the Janko vichian implements. 60The macro mammal remains from the overlying reddish brown loessy loam 61 suggested for an interstadial date, most probably the Szeleta faunal phase (Würm I/II, Hengelo interstadial, MIS 3), noting the striking similarities with the corrected fauna from the Lower layers of the Pilisszántó I rock shelter. 62Later the given layers from these localities were enumerated among the localities of the Istállóskő phase (Denekamp interstadial MIS 3). 63y and large, the stratigraphy of the Pilisszántó rock shelter II is rather problematic: the layer from which the lithics were reported has not been documented and between the two lowermost layers a stratigraphic hiatus is indicated. 64It seems to be evident, that the two artefacts are dated to a period not younger than the MIS 3. In the assemblage of this rock shelter, however, no bifacial implements were found and in our view, the presence of not typical tools and fragments ('raclettes') does not justify the Jankovichian classification of the little lithic assemblage.

Bivak cave
This cave is lying at a distance of 3.5 km from the Pilisszántó rock shelters, in the northern part of the Pilis hill, opening at a large relative height into western direction.During the autumn of 1953 D. Jánossy and L. Vértes excavated the Pleistocene layer sequence of yellow or locally orangecoloured, yellowish grey, grey and brown loam underlying the Holocene humus layer. 65The field works yielded four lithics and two antler tools (Table 1).The bi facially worked leaf shaped tool made of patinated brown radiolarite (nr. 1 in Table 1, Fig. 3.1) and a unilaterally retouched blade were excavated (nr.6 in Table 1, Fig. 3.4) at the border of the grey and the yellowish grey layer.However, the stratigraphic evaluation of the site, similarly to the Pilisszántó rock shelter II showed an important hiatus between these two layers.A 'raclette' of Szentgáltype (nr. 3 in Table 1, Fig. 3.3) and a retouched bladelike flake of greenish grey radiolarite (nr. 1 in Table 1, Fig. 3.2), both made on blanks removed from core edge, were documented in the grey layer, close to each other.Finally, one of the antler tool fragments was most probably found at the border of the orangecoloured and the greyish yellow layer (nr. 2 in Table 1), the other one in the disturbed part of the cave, where only the grey and the reddish brown layers were preserved. 66ncerning the chronology of the grey layer and the lithics, V. GáboriCsánk supposed that the associated fauna is typical to the Early Würm period. 67However, already in the seventies the palaeontologist D. Jánossy enu merated this site among the localities of the Istállóskő faunal phase, 68 noting that there are important differences in the faunal composition of the Bükk and the Pilis sites.Even if this opinion could have been influenced by the pres ence of the 'Szeletian' artefacts, Jánossy pointed out that the remains identified as Megaloceros belong in fact to 60 dobosi-Vörös 1986, 43.-Earlier D. Jánossy dated the reddish layers from the rock shelter to the Istállóskő faunal phase after the presence of Szeletian tools: Jánossy 1977, 143.

Remete Upper cave
This little cave is opening in southwestern direction, at the height of 70 m above the bottom of the narrow and deep Remete gorge (Fig. 4), in the northwestern part of Budapest.The site was excavated by V. GáboriCsánk in 1969-71. 72According to the stratigraphic evaluation of the cave infilling by the geographer F. Schweitzer 73 the large limestone fragments lying directly on the bedrock were dated to the glacial maximum of the penultimate glacial (MIS 6) and the imbedding reddish brown loam was formed during a later mild period (MIS 5e?).The lower part of the overlying layer 4 (reddish loam without limestone fragments) dated to the end of the last interglacial was formed on a discordant surface, showing a washout in the layer sequence.The upper level of the same layer with the lithic artefacts, as well as the remains of 24 vertebrate taxa and the human teeth is yellowish loessy sediment mixed by sharp limestone fragments, documenting the cooling period of the Early Würm (preceding MIS 4).The grey coloured layer 3 with sharp limestone fragments was interpreted as a cryoturbated sediment, 74 overlain by the brown humic layer 2, missing from the inner chamber of the cave and by a recent rendzina soil layer 1 both dated to the Holocene.
The number of the lithics excavated in the upper level of layer 4 was 14 75 or 12, 76 actually eleven pieces are catalogued in the collection of the Budapest History Museum. 77This way, the Remete Upper cave is the second richest locality of the Jankovichian industry in Hungary.Regrettably, in spite of the reports by the excavator 78  72 Gábori-Csánk 1983, 258-263;Gábori-Csánk 1984, 8-10;Gábori-Csánk 1993, 58-60. 73 However, in this layer fossil and subfossil bones and rather atypical Prehistoric sherds were also found, showing that the redeposition dates at least partly to the Holocene: Gábori-Csánk 1983, 253;Gábori-Csánk 1984, 6. 74 Gábori 1981, 98. 75 Including nine formal tools and three flakes: Gábori-Csánk 1983, 267; Gábori-Csánk 1984, 12; Gábori-Csánk 1993, 143.recovery of the lithic artefacts could be unambiguously identified only in four cases.The data enumerated in Table 2 are based on the laconic notes found in the field diary, 79 summarising very shortly the observations of one or two days.
The first lithic, a proximal fragment of a bladelike flake of radiolarite with pebble cortex and very slight traces of secondary modifications (Fig. 5.3) was found in trench 3. 80 A half finished bifacial tool made on a flake of nummulithic chert 81 with dihedral base and centripetal dorsal scars (nr. 9 in Table 2, Fig. 5.1) and a side scraper of Szentgáltype radiolarite with facetted base and unidirectional dorsal scars (nr.8 in Table 2, Fig. 5.2) were excavated at the northern wall of the cave. 82These later tools were reported to be found close to the human remains. 83inally, among the noncatalogued artefacts from the cave there is a box with the label '70.7.21' 84 and '3/1 under the humus, in a yellowish lens'.We suspect that the atypical flake of radiolarite (probably the missing twelfth lithic artefact mentioned by the excavator, 85 the piece of nr. 13 in Table 2) and the two cave bear teeth together with the Glycymeris obovata shell could have been excavated in 1970 in the inner chamber of the cave. 86he details of the recovery of three bifacial tools, which could have been easy to recognise on the field and a half made piece were not documented in the field diary.One of them, a proximal and medial fragment of a long leaf shaped tool of hornstone 87 or poor quality limnic quartzite (Fig. 6.2) was claimed to be one of the typical forms 76 Under the inventory number of 71.1.111.We are grate ful for the friendly help for the colleagues working in the Aquincum Museum, Budapest.
87 Gábori-Csánk 1983, 267, Fig. 16,3; Gábori-Csánk  1984, 12, Fig. 16  of the industry. 88Another piece, a planoconvex leaf shaped tool made on a flat cortical flake with a thinned bulb of percussion (Fig. 7.1) is comparable to the pieces known from Hont lying in the Ipoly valley in northern Hun gary. 89The third piece, a planoconvex leaf shaped tool made of poor quality limnic quartzite is similar in outline to the Moravanytype points, and to a unifacially manufactured piece from the same site. 90The tool from the Re mete Upper cave was thinned on the ventral face by some flat removals with the exception of the tip, where it was inten sively retouched (Fig. 6.1).Finally a piece made of low quality nummulithic chert pebble is interpreted as a half made bifacial tool, probably of a leaf shaped scraper, abandoned by the angular breakage pattern and the hinge re movals (Fig. 7.2).
One of the unifacially worked artefacts found at an unknown place in the cave is a notched tool made on a debordant flake (Fig. 7.4).Its raw material, the brown hydrothermal rock is similar to the pieces known from the Börzsöny Mountains. 91Finally, the plunging flake of dull greycoloured siliceous pebble with a retouched edge was probably fragmented during the excavations (Fig. 7.3).
Among the artefacts there is an amorphous piece with several scars, made of glass, most probably of arti ficial origin.This piece is possibly identical with the 'amorphous volcanic rock' mentioned in the papers 92 or the notched tool of greenish Triassic flint of number 7 of Table 2.
It is interesting to note the presence of the obsidian artefacts (number 2 and 3 in Table 2, found in the topmost horizon of layer 4 in trench 4), which are not mentioned in the publications and cannot be identified in the Palaeolithic collection.In a box containing the Prehistoric artefacts from the cave, however, there is a blade frag ment of Slovakian obsidian and another one of black coloured siliceous rock (which is, however, not obsidian), 93 dated probably to the Copper Age (Ludanice culture), represented among the finds from this site. 94uring the excavations the pit with the Bronze Age depot find was meticulously documented, 95 but no other postPalaeolithic features were described or depicted in the documentation.On the photograph from the inner chamber of the cave, however, a section of a pit is clearly visible (Fig. 4), 96 suggesting that some lithics including the obsidian artefacts or the mentioned amorphous piece could have been intrusive finds in the Pleistocene layers.
Unfortunately, the mammal remains from the site have not been systematically analysed yet and only preliminary data are available from layer 4 of the first chamber (data by the palaentologist M. Kretzoi). 97According to the excavator, both the composition of the fauna (including remains of Lagopus, Ursus spealaeus, Crocotta, Equus, Leo, Coelodonta and Megaloceros) found in the upper level of layer 4 and the presence of the muskox Ovibos, is typical for to the period preceding the first Würmian Pleniglacial (Early Würm or 'Altwürm' period, MIS 5da). 98Recently, however, I. Vörös revised the Ovibos bones and concluded that the remains from the Remete Upper cave belong to a little bison 99 and placed the age of yellow layer 4 into the Szeleta faunal phase, i.e. to the Hengelo interstadial. 100Taking into consideration the last appearance date of the giant deer in the Carpathian Basin, which is the discriminating species of the Szeleta and Istállóskő phases or the Hengelo and Denekamp interstadi als, 101 the age of the fauna associated with the lithic tools cannot be securely placed to the first part of the last gla ciation. 10288 E.g.Gábori 1964, 72, Pl.XIX.1; Zandler 2010, Fig. 9.1. 89Zandler 2010, Fig. 9.2.-In the Moravány assemblage excavated by J. Bárta there are similar pieces without flat ventral re touch, too: neMerGut 2010, 192. 90The same raw material is also known from the Hont as semblage.Moreover, on the piece of the Remete Upper cave a number "124" written by ink is clearly visible.A similar label is found on the artefacts from Hont again.
91 Gábori-Csánk 1984, 12. -In other papers this piece is referred as amorphous block of volcanic rock and it was suspected that the source region was not in the Tokaj Mountains: Gábori-Csánk 1983, 267;Gábori-Csánk 1993, 143. 92 Stored under the inventory number 73.Charcoals remains, exclusively belonging to larch or spruce (Larix-Picea group) were reported from layer 4 in the entrance of the cave, which seemingly confirmed the age of the layer and the human occupation predating the first Pleniglacial.104However, pieces of charcoal were also found in the inner chamber in 1970105 and according to the field diary further fragments were observed e.g. in layer 4 of trench 3, where the first artefact (Fig. 4.3)106 was found.Moreover, the report on the analysis of the charcoal pieces written in the Department of Applied Botany and Tissue Evolution of the Eötvös University, Budapest, and completed 5 December 1969107 clearly shows that in the samples a number of species, basically of deciduous trees were recognised (Table 3).In fact, larch or spruce was identified only among the charcoal pieces collected from the sediment of the penultimate glaciation and in another one associated with two hornstone flakes (nr.4-5 in Table 2).However, in this latter layer remains of maple, dogwood and common oak were also found.Otherwise, the deciduous trees were common in the other samples, which, together with the presence of the intrusive artefacts may raise the question of later mixing of pieces in the cave sediment. 108fter this short review it is clear that the age of the Palaeolithic assemblage from the Remete Upper cave is a very problematic question: the faunistic material has not been analysed in details and the palaeobotanical data col lected during the first excavation in 1969 and never published before reflects an unexpectedly complex picture.Sev eral observations and indirect data suggest for the disturbance of the sediment, which were, however, not sufficiently documented.Bearing in our mind the low number of the lithic artefacts as well as the questions and doubts concerning their provenance, we conclude that this locality in itself is not adequate to define a distinct archaeological industry.

THE ASSEMBLAGE FROM THE BIVAK AND REMETE UPPER CAVE: JANKOVICHIAN, SZELETIAN OR A LEAF SHAPED INDUSTRY?
The recent review articles on the 'Szeletian'109 reflect a certain dichotomy from Hungary, based on the assemblages from two old excavated sites, the Szeleta and Jankovich cave.However, in 2003-2005 and 2007  The evaluation of the assemblage 110 led to the description of the 'Vanyarc type industry' defined after the typologi cal evaluation of the lithic tools as well as the observation on the use of the raw material types, with the emphasis on the extralocal rocks.Moreover, in the last years rich and well preserved assemblages were also excavated in the same region, at Szécsénke and Galgagyörk. 111In our view, it is important to develop hypothesis and research ques tions from the well documented assemblages, 112 which make possible to raise an issue of the variability of the leaf point industries in Northern Hungary.From this aspect the 'Jankovichian' and 'Szeletian' are less well defined variants of the Middle or Upper Palaeolithic entities with single poorly documented sites.
Both the Szeletian and the Jankovichian industries have been defined on typological ground, i.e. after the presence of leaf shaped implements.As these artefacts are missing from the Pilisszántó rock shelter II, this site cannot belong to these entities. 113At the same time, the interpretation of the recently excavated assemblages from the Dzeravá skála cave pointed to an interesting problem.In the uppermost level of layer 11 a 'raclette' and a small flake from bifacial retouch, 114 in the overlying layer 9 among others a flake from the flat retouching was found. 115he presence of the waste material from the manufacture of bifacial tools raises the question on the '[post]Leafpoint' and 'Aurignacian' classification of the little assemblages, especially, that the associated 37 ka B. P. radiocarbon ages are very similar to the dates published from the Szeletian sites in Moravia.
The Bivak and the Remete Upper caves are lying at a large relative height above the bottom of the valley and most probably served for very short occupations (ʻbivouac siteʼ), 116 similarly to the Istállóskő cave.Accordingly, the number of the excavated artefacts is low and their typological composition is rather onesided: apart from the bifacial implements the formal tools are represented by side scrapers, and ʻSzeleta scrapers' or ʻraclettes', i.e. partly naturally modified pieces, partly reshaped blanks.The Middle Palaeolithic character of the assemblages is also reflected in the flakes with facetted base; however, in the absence of cores the technological evaluation is rather problematic.
Compared these assemblage to that one known from the Jankovich cave, the differences are found in the quantity and not the composition of the tools, suggesting that on this later site a palimpsest of very short occupations could have been excavated.However, the field works at this locality were not documented sufficiently.
It is important to stress the role of the extralocal rocks and pebble raw material used by the humans.In the Bivak cave, the three tools of radiolarite were made on three different macroscopic variants (including the Szent gál type 117 ), one of them certainly of alluvial origin.In the Remete Upper assemblage, beside the artefacts of nummu lithic chert 118 (Fig. 5.1;Fig. 7.2), the tool of grey silex (Fig. 7.4) and at least one of the radiolarite artefacts (Fig. 5.3) 119 were certainly made on pebble raw material.In this later case, the source area of the pebbles is most probably found in the southwestern part of Budapest, lying at a distance of 15 km from the site, where large out crops of the Lower Miocene Budafok formation were found.From the same region sand layers with abundant Glycymeris remains, dated to the Oligocene were also reported, 120 suggesting that the fossil shell could have been transported to the site together with the lithic raw material types.uncovered so far.Our understanding of the Szeletian in Hungary is still based on the archaeological sequences from the two caves' is not correct: Mester 2018, 34;cf. Mester 2014b, 160. 110 Excavations by K. Zandler and A. Markó.The field re ports from these sites are under preparation. 111The average find density of the artefact bearing layer at Vanyarc was 37.5 pieces per square meters, while the maximum num ber of the pieces excavated in the Szeleta cave on 4 square meters in artificial levels of half meter in thickness was 29: Markó 2012, 214-215; Markó 2016, 24-27.-Although Zs.Mester thinks that it is meaningless to use the find density data for the Szeleta assemblages, these values reflect both the low number of the excavated artefacts and the factually low resolution of the documentation from the Szeleta, compared to th recent excavations: Mester 2018, 35.
Gábori-Csánk 1983, 265;Gábori-Csánk 1984, 11. 117 This term has been used since 1984 for a characteristic macroscopic variant, named after its most important outcrop.How ever, beside Szentgál a number of occurrences of this type were re ported from the Bakony mountains (Lókút, Hárskút and Bakony csernye) and even from Gerecse (Pisznice), lying at a distance of 20-25 km from the Bivak cave and 35-40 km from the Remete Upper cave: bíró 1984, 49.
118 At this raw material, the raw material of the artefacts without cortical surface are identified collected from secondary sources, as the primary outcrops of this special siliceous rock are not known.The occurrencces of nummulithic chert pebbles were reported from several pebble bearing formations dated from the Oligocene/ Miocene to the Holocene: Markó-káZMér 2004.
119 Earlier GáboriCsánk has taken into consideration the primary radiolarite outcrops around Dorog: Gábori-Csánk 1983, 269;Gábori-Csánk 1984, 13. 120 FöldVáry 1929, 38-40.Earlier we compared the Remete Upper cave assemblage to the uppermost artefacts from the Jankovich cave and to the Szeletian industry, after the presence of the pieces similar to the Moravány types. 121n a recent paper, however, we argued, that the name of this cultural entity is inappropriate, as it was not possible to define a welldefined assemblage from the eponymous site as the typical Szeletian material.Moreover, because of the problems with the excavations and the documentation of the Szeleta cave, as well as the questions concerning the site formation, we suggested to use 'Szeletian' exclusively for the localities of the Bükk Mountains. 122oncerning the Jankovichian, one of the few characteristic pieces of the Jankovich cave are the Levallois flakes and points thinned on their ventral face. 123Recently Zs.Mester, disputing this observation drew attention to the presence of a similar tool in the Háromkút cave (Fig. 8.1) in the Bükk Mountains. 124Though not mentioned, Mester seemingly follows the view by J. Hillebrand by enumerating the Háromkút tool among the 'Protosolutréan' finds. 125However, as similar pieces are not found in the in the Szeleta cave, the tool of the Háromkút cave is gener ally compared to the artefacts of the Transdanubian Szeletian or the Jankovichian and the Middle Palaeolithic 'rolled' industry of the Gudenus cave in Lower Austria. 126. Gábori was the first to mention from the same site a 'D shaped scraperʼ 127 or raclette made of Korlát type material, also used in the Szeleta and Istállóskő cave (Fig. 8.2).Finally, according to the inventory book a less typical tool with alternating manufacture made on a tabular piece of metarhyolite/felsitic porphyry (Fig. 8.3) be longs to the same assemblage. 128These later artefacts were suspected to be collected by Hillebrand in 1923, 129 however, according to the excavator, during this season no archaeological artefacts were found. 130Apart from the uncertainty concerning the circumstances of the recovery of these artefacts, none of them can be regarded as a particularly typical form for the Early Szeletian of the Szeleta cave 131 and this way, the assemblage of the Háromkút cave does not belong to this entity.
On the other hand, the Levallois blanks with ventral thinning at the tip of the tools were recently interpreted as half made tools. 132The scars on the artefact of nummulithic chert from the Remete Upper cave (Fig. 4.1), 133 how ever, shows an entirely different manufacturing strategy, starting with broad removals from the whole perimeter of the ventral side of the flake.Furthermore, ventral thinning at the tip of another bifacial tool from the same site, similar in its outline to the Moravány types 134 (Fig. 5.1) was clearly posterior to the large flat removals covering the entire ventral face.
In any cases, Levallois points and flakes with flat fine ventral thinning at the tip of the tools at their distal part, similar to the Jankovich and Háromkút items are missing from the studied assemblages.This way, the Janko vichian classification of Bivak and the Remete Felső caves is rather problematic.
In this paper several pieces were mentioned from the Remete Upper cave showing apparent typological similarities with the tools from HontCsitár or which were made on the raw materials known from the northern part of the Börzsöny Mountains.In 1964 the openair site of Hont was discussed among the Mesolithic localities. 135Five years later M. Gábori performed excavations on the site, but regrettably, no documentation is accessible from the fieldworks 136 and in 1976 an industry with both Middle and Upper Palaeolithic elements was shortly mentioned from Hont; 137 the Upper Palaeolithic and later artefacts were handed to the Palaeolithic collection of the Hungarian Na 121 Markó 2013a, 20.
135 Gábori 1964, 70-72. 136 These excavations are not mentioned even among the yearly short reports on the archaeological research, listed in the vol ume 97 of the periodical Archaeologiai Értesítő.See: Gábori 1981, 100, note 13. 137 Gábori 1976, 80-81.-cf.Gábori 1964, 70-72.tional Museum in 1973. 138In connection with the Jankovichian, HontCsitár was mentioned along with other as semblages from the openair sites in Northern Hungary, 139 but the bifacially worked and Middle Palaeolithic artefacts were added to the Palaeolithic collection only after the death of the excavator in 1996.Regrettably, some tools, published in 1964 140 were found together in the boxes with the lithics of the 1969 excavations, indicating the mixing of the artefacts. 141Luckily, the site was successfully reidentified in 2002 by K. Zandler, A. Péntek and A. Markó.This way, in the future it will be possible to organise control excavations on this important site with the aim of clear ing the stratigraphy and the chronology of the artefact bearing layer, as well as the typological composition of the lithic assemblage.On the other hand, some new data concerning the cultural classification not only of this site but indirectly also of the Remete Upper cave assemblage is expected.

QUESTIONS OF THE CHRONOLOGY
The Early Würmian age of the Jankovichian industry, proposed by GáboriCsánk was based on the mam mal biostratigraphy, namely the composition of the faunal assemblages.The recent paleontological evaluations, however, suggest for a more recent age for the layers discussed in this paper, even if the systematic analysis of the faunal assemblage from the Remete Upper and probably the Bivak cave, too, may provide a new biochronological date of the sites.
The archaeological classification of the ʻJankovichianʼ industry is rather problematic.In 2017 M. Patou Mathis and her colleagues seemingly refuting our idea pointed out that the lower layer of the Istállóskő cave does not belong to the Late Middle Palaeolithic Jankovichian industry. 142In the same year, one of the coauthors of the same paper, virtually arguing against our observation on the presence of the Moravány type leaf shaped tool in the Jankovich cave, 143 emphasised that in the Moravány-Dlhá assemblage blades and blade cores, missing from the Jankovich assemblage, are also represented. 144However, on the same page Mester accepted that the artefacts found in the upper levels of the Jankovich cave can be dated to the Early Upper Palaeolithic period. 145inally, in a paper published in 2016, three coauthors by M. PatouMathis reported the presence of Late Gravettian artefacts from the Szeleta cave.Although in layer 5 asymmetric leaf shaped points attributed to the Jankovichian or Early Szeletian industry were also found, the conclusion, that this assemblage '...can represent a Late Gravettian with leaf pointsʼ, 146 suggests that 'Jankovichian' lithics in this layer of the Szeleta cave are found in association with the tools dated to the Middle Upper Palaeolithic.
On the other hand, without attaching great importance of the 13 bifacial pieces excavated in more than 950 cubic meters of sediment in the Szeleta cave, after excluding the bifacial tools wearing pseudoretouch from the study the central hypothesis by Mester was refuted in this paper, as the majority of the asymmetric leaf shaped tools were found in the upper layers 5 147 and 6. 148 This way, these pieces are neither characteristic forms of the lower 138 Including some earlier published pieces: Gábori 1964, Fig. XIX.7-14,17. 139 Gábori-Csánk 1983, 285;Gábori-Csánk 1984, 21. 140 Gábori 1964, Fig. XIX.1-4,15-16;cf. Zandler 2010, Fig. 8.3-5, Fig. 9.1-2. 141 In fact, the problems with the site and the assemblage started earlier: the locality, first mentioned under the name of Hont Csitár is in fact identical with HontBabat following the present termi nology, but not identical with the site excavated by M. Gábori: Gábori-Csánk 1958, 60-61, obr. 35;Zandler 2010, Fig. 33,1. 142 Patou-Mathis et al. 2017, 86.-Otherwise, this is a typical example for the straw man argument, as we never classified the given assemblage as Jankovichian: Markó 2017, 194, fn. 14, 15. 143 This was, however, suggested by J. Bárta more than fifty years ago : bárta 1960, 310 -cf. hillebrand 1926. 144 Mester 2017, 86.-In our view, the validity of the ty pological classification of a single artefact cannot be questioned or refuted by technological arguments based on another much larger as semblage.Moreover, the differences in the site function, i.e. hunting stand in a cave versus openair workshop site makes the comparison of the assemblages rather difficult : neMerGut 2010, 190-191, 193. 145 ʻPour chaque gisement [i.e. both the Jankovich and the Szeleta cave], les pièces de la partie inférieur de la séquence peuvent appartenire à une industrie de Paléolithique moyen (tardif?), tandis que celles de la partie supérieure peuvent être d'âge du Paléolithique supérieur ancien.ʼ:lenGyel et al. 2016, 181. 147 At least in the side corridor of the Szeleta cave, where artefacts from layer 5 were documented, no important sediment mix ture was observed.In our view, the composition of this assemblage undermines the scientific value of the results of the 'technological approach' in the case of the Szeleta and Jankovich caves: Markó 2016, 24, 27-31. 148 lenGyel et al. 2016, 177, Table 5. -As a matter of fact, the numbers of leafpoints given in this table differ from those found in Table 1 and 4 of the same paper, cf.note 40.layers of the Szeleta cave nor do they belong to the Early Szeletian and/or Jankovichian industry.Moreover, fol lowing the logic used at the Dzeravá skála assemblage, the 4446 ka B.P. radiocarbon dates from layer 2 of the Szeleta 149 cave may imply a Micoquian affiliation, 150 showing that the asymmetric leaf shaped tools are equally known from the Middle, Early Upper and Middle Upper Palaeolithic layers, and these pieces are not exclusively typical for the Jankovichian and/or Early Szeletian entities.
Earlier we emphasised that in lower layer of the Istállóskő cave, 151 as well as in the Jankovich and Bivak caves the Middle Palaeolithic type and bifacially manufactured lithic tools are associated with the Aurignaciantype osseous industry, 152 presenting the typical case of the ʻobservational contemporaneityʼ. 153Without giving a compre hensive review of the Central European and Balkan sites 154 we refer to the recently published 36 ka BP radiocarbon age 155 from layer XI of the Obłazowa cave in the Polish Carpathians, where both the osseous point and refitted fragments of a bifacially manufactured leaf point were excavated in the same layer, relatively far from the disturbed part of 'layer XXII', 156 i.e. in an in situ position.
Finally, the 35.6 ka B.P. direct date of the chisel of antler from the Pálffy/Dzeravá skála cave 157 is slightly older than the previously published 3131.6 ka B.P. dates of the points 158 from the same site and it seems to be consist ent with the above mentioned ages of layer 9 and the uppermost level of layer 11, where bifacial technology is repre sented by characteristic waste material and a fragment of an osseous point was also excavated. 159In our view, the osseous artefacts from the Bivak as well as the Jankovich cave will provide similar radiocarbon ages, dating the os seous artefacts to the Middle Würm, approximately to the same period as the Aurignacian sites.

CONCLUSIONS
In the 21 st century the importance of the Szeleta and the Jankovich caves is found in their role of the Pa laeolithic research in Hungary.The history of these sites and the ʻSzeletianʼ and ʻJankovichianʼ industries clearly show that during the evaluation of the results, the data both from the technical literature and the unpublished re search reports were often used selectively.Moreover, instead of systematic analysis, ad hoc arguments, informal fallacies and misstatements were also used during the reasoning.Reasoning and important differences in both the numerical data and the interpretation of the same assemblage are frequently noticed in the papers written by the same authors.The most important problem is, however, that the theories and the general reviews are based on the analysis of single pieces and not on the assemblages. 160n the present paper we found that the Bivak cave is the single 'Jankovichian' locality, which was ade quately documented using the standards of the age of the excavations.The analysis of the faunal assemblage from the Remete Upper cave may give a solid biochronological background of the archaeological artefacts in the future, however, the field observations concerning the lithic tools are largely missing.Finally, in our view, the recent pub 149 hauCk et al. 2016, Tab. 4. 150 kaMinská et al. 2005, 55;hauCk et al. 2016.-Mester seemingly recognised this when he dated the Early Szeletian from the Szeleta cave to 4441 cal.B.P., even if in the same paper he sorted the asymmetric pieces from layer 2 to the Early Szeletian assemblage: Mester 2017, 86, cf.Mester 2017, 78.
151 Mester suggested, that the leaf shaped scraper from the lower ('Aurignacian I') layer of this cave at a distance of 8-10 meters from the trenches where the majority of the bone (in fact: antler) points were documented.This is however, a misleading statement, as in trench VIII lying very close to place of recovery of the bifacial tool five osseous artefacts including two split based points were also exca vated from the lower layer: Markó 2017, 206, 216, 159 During the excavations by F. Prošek more than twenty osseous tools were excavated, clearly from the Szeletian layer (in the lower part of the greenishgrey -brownish grey layers 5-11 following the stratigraphy published in 1951 or layer 8 in 1953): Prošek 1951, 296, 297-298;Prošek 1953, 185, 191. 160 The typical examples are the technological investiga tions of the leaf shaped implements, which, ultimately illustrate the return of the dichotomy of the leaf shaped implements on one hand and the accompanying industry (Begleitindustrie) on the other one.This rightly criticized approach was typical in the 20 th century, see: Mester 2018, 21 -In agreement with the view expressed earlier by Mester and rejected later without proper argumentation, the leaf shaped implements in themselves are not suitable for archaeological classification: Mester 2011, 29; Mester 2014a, 53.lication on the artefacts from the Dzeravá skála cave allows supposing a leafpoint assemblage (ʻSzeletianʼ) in layer 9 and 11.
Regrettably, the artefacts from the Jankovich cave were excavated almost exclusively without stratigraphic control and some pieces were erroneously attributed to this assemblage. 161Besides, a few pieces, collected in some cases as stray finds or without stratigraphic observation and having questionable cultural classification have been incorporated to the Jankovichian.This way, after the review of the assemblages excavated after the World War II in Hungary, it is clear that it was prematurely defined as a separate cultural entity.On the other hand, the use of the term 'Szeletian' should be restricted for the Bükk sites, 162 where the occurrences of the asymmetrical leaf shaped tools are independent from the chronological or cultural classification of the similar pieces of the Jankovich cave.
From our part, we suggest to the use the term ʻleaf shaped industryʼ (Blattspitzenindustrie) 163 for the di verse lithic assemblages with Middle Palaeolithic elements from the cave sites in the Transdanubia and in the Bükk Mountains, 164 as well as for the recently excavated assemblages in the Cserhát Mountains.Beside the new excava tions of the openair sites and possibly, in the Istállóskő cave 165 the next step in this research can be the investigation of the 'modern traits' known from the assemblages: the intense use of the extralocal raw materials on the Cserhát sites and the presence of the osseous industry in the cave localities.

Fig. 1 .
Fig. 1.The localities discussed and mentioned in this paper (map constructed by B. Holl) 59

Fig. 3 .
Fig. 3. Bivak cave: lithic artefacts (drawing by K. Nagy) and the distribution of the including the antler tools (displayed by stars) in the cave (following Jánossy et al. 1957 modified)

Fig. 4 .
Fig. 4. Remete gorge: the entrances of the caves during the excavations and the longitudinal section in the inner chamber of the Remete Upper cave 69

Fig. 7 .
Fig. 7. Remete Upper cave: leaf shaped implement, half made tool and retouched pieces from unknown parts of the cave (drawing by K. Nagy)

Table 3 .
1950 lithic artefacts including 32 typical tools and 10 retonched fragments of a Middle Palaeolithic industry with leaf shaped implements were excavated in a not disturbed artefactbearing layer near Vanyarc, in the Cserhát Mountains.Charcoal remains from the 1969 excavaions of the Remete Upper layer