The arm-shaped vessels in Anatolia and the Eastern Mediterranean during the Late Bronze Age: a morphological and contextual analysis

Abstract. The so-called libation arms found in Anatolia and the Eastern Mediterranean during the Late Bronze Age, belong to the ceramic class Red Lustrous Wheel-made Ware, characterized by a very fine fabric, a careful cooking and a red slipped and polished surface. Even if they were thoroughly analyzed, it was generally assumed that these objects were linked to religious or cultic activities and destined to libatory action. However, no systematic investigation was carried out in relation to their finding contexts. This paper presents the results of a morphological and contextual analysis of this specific artifact. It offers suggestions for production areas, function and distribution on the base of a catalog that collects all the pieces found so far and on the analysis of each finding context. The data seem to indicate an Anatolian type of production unrelated to that of the Red Lustrous Wheelmade Ware, which is solely linked to religious activity.


INTRODUCTION 1
The aim of this paper is to present typological and contextual observations on, as well as hypothesize about the distribution of, the so-called libation arms. Libation arms are vessels in the shape of an outstretched human arm ending in a hand holding a cup that were distributed not only in Anatolia, but also in Cilicia, the Levant and Cyprus -a large area that, during the Late Bronze Age, experienced a period of intense economic, political and social interactions. This research includes the study of the published material, taking into consideration all the sites where the libation arms were found that have already been mentioned in the study by K. Eriksson (1993) and updated by D.P. Mielke (2006: Table 2), and adds the most recent finds in other sites not mentioned in the two previous studies. Furthermore, this research also focuses on the analysis of the contexts of these specific vessels 2 . This study led to the creation of a database to collect all the published libation arms and updated distribution maps, as well as established a first typology and a possible correlation between the morphological type and the context. The single contexts have been analysed in detail and reduced to five main categories based on the information reported in the excavation reports. Since it is not possible to explain the individual contexts in detail here, the context category to which each artifact belongs is listed in the appendix. At the same time, a systematic catalogue was compiled to collect all published libation arms 3 . In addition, the research and the database were implemented through the study, still on-going, of the unpublished libation arms 4 recovered from the Southern Ponds secondary filling of Boğazköy/Hattusa, partially analysed by T. Pilavci (2017) in her PhD thesis 5 and not discussed in this article. Therefore, the goal of this research is first to analyse the libation arms by focusing on their morphological features, which are studied only marginally in the literature. This research helps demonstrate the presence of workshops in certain areas by shedding light on the function of the libation arms -different morphological types of libation arm may could belong to different production centres that may further be associated with a certain style. Secondly, this study focuses on the geographic and chronological distribution of libation arms to combine the typology and workshop analysis with interregional connections. The article also takes into consideration the finding contexts to better define function and chronological distribution.

CRAFT PRODUCTION
The so-called libation arms ( Fig. 1) consist of three parts manufactured separately: a long cylindrical tube, in most cases wheel-made; a hand-modelled part that includes the hand and the bowl it supports; and a third element that is the junction between the two parts, represented by modelled rings that can differ in number (from two to four). The libation arms are closed vessels that are hollow inside and whose only opening is a perforation visible on the side of the bowl. Most of the finds belong to the Red Lustrous Wheel-made Ware (RLWmW 6 hereafter) class, which is characterised by a compact fabric, consisting of a very fine red or orange clay with few inclusions and uniformly fired. The surface, in most cases, is covered with a red or orange self-slip and is carefully polished to give a shiny appearance. The vessel's dimensions are variable -some examples are longer than 60 cm, while others, according to the diameter of the bowl, appear to be much smaller. The hand that is represented is the right hand. The bowl, in almost all cases, is deep and in a few cases is wider and shallower. The bowl is supported from the bottom, with the fingers sometimes wrapping around it up to the edge of the rim, although in two examples the hand is holding the bowl from the side. Table 1 presents the distribution of the arm-shaped vessels according to the dating of the contexts provided by the archaeologists who published them, considering that the artefact could be dated to a period previous or contemporary to the context itself. Until a reconsideration of the dating of these contexts takes place, the dates are considered here as valid. According to the data collected during the Master thesis, it appears that the 'oldest' sherds, belonging to the 16 th -15 th centuries BC, 9 are mostly found in public contexts (palace, administrative, temple) and the only private context is the funerary one. During the 14 th century BC there is no evidence of arm shaped vessels from funerary contexts (mainly due both to the lack of Late Bronze Age burial sites in Anatolia and to the impossibility of dating some of the finds in Cypriot burials), and all the vessels are retrieved from domestic/ templar contexts until the 13 th century BC, when almost all the sherds are recovered from temple contexts.

CONTEXTS AND THEIR DATINGS
The type of context (cf. Tab. 1) with the highest number of sherds is the temple context type (or its immediate surroundings). This could support the hypothesis of a cultic function of the object. The fact that the sherds recovered in this type of context come not only from Anatolia (which generally has the largest number of specimens) but also from Cilicia and Cyprus may suggest that their use is indeed cult-related. The date of these vessels seems to not extend beyond the 14 th century BC, also considering the uncertainly of some Cypriot contexts. The sherds from Cilicia for which a dating was provided, like those from Kilise Tepe, cover a time span between 15 th and the second half of the 14 th century BC (Kibaroğlu et al. 2019: 415). The finds from Anatolia cover a time span from the 16th to the end of the 13 th century BC, while in the Northern Levant the libation arms seem to appear later, during the 14 th -13 th centuries BC, as the arms from Tell Atchana and Ugarit shows.

PROPOSED TYPOLOGY
The typologically relevant parts of a libation arm are the hand, the wrist, and the base, the first two of which are connected to the third by a wheel-made tube (which can be cylindrical or slightly wider towards the base, mainly ranging between 50 and 70 cm in length but without other useful typological characteristics). The rest of the object is hand-made. Until now, no specific morphological analysis has ever been carried out on the rendering of the hand, the rings decorating the wrist, or the base in order to establish a typology for the libation arms. In previous studies, the only morphological distinction was related only to the size of the whole object: a specimen from Enkomi (Cyprus), complete with base, wrist, and hand, has for a long time led to the assumption of the existence of a long and a short type of libation arm (Bittel 1957: 33-42). Pilavcı (2017: 116-117) makes a distinction between long and short types, adding two new types: miniature and votive. She distinguishes the characteristic parts (fingers, wrist, and base) but makes morphological distinction only for very noticeable exceptions. In this study, I decided not to refer to this type of dimensional distinction introduced for the first time by Bittel (1957: 36-38) and based on the length of the object because only nine vessels out of 220 analysed and catalogued can be defined as complete: eight would belong to the 'long' type while only the Enkomi specimen would belong to the short one. Although the state of preservation of the remaining libation arms is fragmentary, following observations on the tube fragments can help in solve the question long/short arm, it seems to remain constant throughout the preserved length in most cases, while the diameter of the specimen of libation arm defined as belonging to the short type increases visibly towards the base. None of the fragments analysed that are part of the arm show an accentuated increase in base diameter that would make them part of the short type. Therefore, it can be assumed that the long type is the most common in the analysed areas. The morphology presented here is related to the single parts forming the vessel rather than to its general shape.

Base
Six base types could be identified among the 50 bases recorded: ring base (BA, Fig. 13), narrow ring base (BAS, Fig. 14), disc base (BDI, Fig. 15), button base (BB, Fig. 5), flat base (BP, Fig. 16) and rounded base (BR, Fig. 17).The majority of the bases belongs to the narrow ring and button type (9 and 8 sherds each respectively) and suggests that the primary function of these objects was not to be used in an upright position, as the vessels do not show a suitable base for this purpose.

Wrist
The wrist is defined as the junction between the wheel-made part of the arm and the handmade part of the hand, a point that corresponds to the anatomical part of the human wrist. The joint is highlighted through a decoration commonly found on all the wrist fragments. It consists of a series of 'rings' of different sizes and executed in different ways. In most cases they are in relief and are placed close together, but it cannot be excluded the existence of examples with an incised decoration along the circumference of the wrist and other rings in relief with some space between one and the other. Among the 66 sherds with complete or partial wrist decoration, it was possible to distinguish three types of decorative execution of the rings: relief decoration (PRI, Fig. 18), spaced relief decoration (PRID, Fig. 7), incised decoration (PIN, Fig. 8). The analysis of the number of rings executed on the vessels allowed the definition of vessels with three rings, two rings and four or more rings.
These data show that the most common combination between the number of rings and the way they are represented is three rings in relief. There does not seem to be any correlation between the number of rings and the way they are executed: one can find three incised rings as well as two raised rings without any apparent precise pattern.

Hand
The hand, together with the bowl, represents the frontal part of the libation arm and the only way to fill the container.
The analysis of 80 sherds displaying the hand at various degree of preservation led to the definition of some distinctive criteria to group them. These are based on the rendering of the thumb, fingers, and nails 10 .
The subdivision according to the morphological differences led to the creation of ten groups: M1 ( Fig. 19): Thumb in high relief, fingers of the same length, even, stop before the rim, executed in low relief and with naturalistic nails; M1a (   -15 th century BC  7  11  5  9  5  37  14 th century BC  9  3  33  3  48  13 th -12 th century BC  52  6  3  1  62  ND  10  1  37+31  66  TOTAL  68  9  24  33  9  10  5  68 Thumb in high relief, fingers of the same length, even, stop before the edge, naturalistic renderings and naturalistic nails; M2a ( Fig. 9): Unique variant of category M2, the whole hand is realistically rendered, the fingers are wellspaced and the knuckles are recognisable. M3 ( Fig. 21): High-relief thumb, fingers of the same length, even, stop before the brim, rendered in low relief and with carelessly incised nails; M3a ( The largest group is M3, followed by M2, M4 and M1, excluding groups M1a and M9 which are mostly composed of fragments that are too poorly preserved.

Production areas 11
Based on the morphological analysis, it is evident that the most discriminating element defining the shape of the libation arms is the hand, identifiable on approximately 36% of the pieces compared to the total number of analysed sherds (220), while typology of the bases and wrists did not show relevant results. The following four main types of hand may refer to different workshops.
Group M3, which can be summarised as a poorly made hand, is mostly found in Boğazköy, both in the Lower Town (B39; B42) dating to the 14 th century and in the Upper Town (B52; B55; B68) and near Temple 15 (B64) dating between the 13 th and 12 th centuries BC. (Fischer 1963: 149-150;Parzinger-Sanz 1992: 116). Two arms from Boğazköy (B02; B03) come from a layer without any information on context or dating (Bittel 1937: Table 16). The remaining libation arms come from Ortaköy-Šapinuwa (OS13; OS14), more specifically from the remains of Building D, dated between the 15 th and 14 th centuries BC (Kiymet and Süel 1999: 474); from the Late Bronze Age Levels of Korucutepe (K07) (Ertem 1988: 18) and Kilise Tepe (KI04) (Symington 2001: 169-170); from Tomb 2 of Enkomi (EN04) dated to the 14 th century BC (Courtois et al. 1986: 18, 27-28); and from Level 2 (Hittite architectural level) of Alaca Höyük (A02) (Koşay-Akok 1966: 169). The libation arm YU01, coming from the level of the early 15 th century BC of Yumuktepe, differs macroscopically from the other vessels: the surface, instead of being red, is closer to brown and the body is dark brown and richer in inclusions than the typical RLWmW body. It has been suggested that this is a local production, based also on the evidence of a less careful manufacture if compared to the other specimens (Manuelli 2009: 259-260). The same brown-coloured clay can also be found, however, in fragment B40, recovered from Level 1 (14 th century BC) of the Lower Town of Boğazköy.
The second largest group, M2, is characterised by a more naturalistic execution of the hand. Among the thirteen vessels in this group, four come from Cyprus: AP01, for which no date has been given (Ohnefalsch-Richter 1893:385); AI02 from the 14 th century BC sanctuary (Gjerstad 1934: 358); EN02 from Tomb 69 dated between the 15th and 14 th centuries BC (Åström 1967: 8;Courtois et al. 1986: 41); and MR01 from Tomb 7 dated between the 15 th and 13 th centuries BC (Åström 1972a: 205). The fragments from Boğazköy were mostly found in the Lower Town (B19; B37) and one of them (B50) is the only specimen recovered not far away from Temple 1. Only one vessel (B96) comes from the Upper Town, from Temple 6, while a single sherd (B07) comes from the citadel of Büyükkale (Fischer 1963: 150). Fragment UG01, on the other hand, belongs to the Ugarit Recent 2 period, corresponding to a time spanning from the second half of the 15 th century BC to the first half of the 14 th century BC (Schaeffer 1949: 210). Of the specimens from Alaca Höyük, one sherd (A03) could not be contextualised but belongs to the Hittite period level, while another sherd (A10) is classified as belonging to Level 2 (Koşay-Akok 1966: 169). Only one vessel belonging to this group comes from Korucutepe (K06), found in L-17 I.tb 400-410 but there is no date (Ertem 1988: 18). There are no substantial differences in the ware, which is fine and reddish, nor in the reddish-orange and polished surface.
The arms of group M4, characterised by a poor execution and the fingers converging towards the centre, were all found in Boğazköy and all within the Upper Town Area, dated between the 13 th and 12 th century BC: fragments B75 and B77 belong to Temple 15; fragment B72 came to light south of Temple 4; and fragments B57, B78 and B82 were found in the debris without any context. All sherds share the same type of fine, reddish body and reddish-orange polished surface (Parzinger-Sanz 1992: 116). One fragment recovered from Kaman-Kalehöyük also belongs to this group, but it is not possible to define the context of its discovery (Omura 1999: 219).
The fragments belonging to group M1 may, at first glance look the same as group M3, but they are distinguished by a more naturalistic rendering of the fingers and nails. Fragments B53 and B84 come from the Upper Town of Boğazköy, found near Temple 4 and fragment B76 from Temple 15 (Parzinger-Sanz 1992: 116), all dated between 13 th and 12 th century BC; the libation arm AA01 was found inside House 37 in Alalakh dating from the mid-15 th century to the first half of the 14 th century BC (Woolley 1955: 178); and fragment KA04 comes from Building B in Kayalıpınar, dated between the 15 th century BC and the 14 th century BC, which has been interpreted as an administrative building (Mühlenbruch 2014: 115-117). The sherds from Boğazköy have the same fine, reddish ware and reddish-orange polished surface. The surface of the fragment from Alalakh is also red and polished, while the fragment from Kayalıpınar cannot be described more precisely.
The M7 group, although it includes few sherds, is one of the most interesting. The hand differs visibly from that of the other groups as its rendering is rather naturalistic, even if the knuckles are represented with a clear detachment from what should be the back of the hand and the fingers are separated by deep incisions that make them appear in relief. The thumb is still made in relief, but the last phalanx is very prominent compared to the other vessels. The specimen from Enkomi, EN01 (Courtois et al. 1986: 44-45), was found inside Tomb 57, in use from LC I to LC II A-C (17th-12th c. BC (Steiner and Killebrew 2014: Tab. 4. 3). It has been suggested that the fragment from the Dede Mezarı Necropolis (Üyümez et al.: 2010: 939-943, 949) belongs to the Middle Bronze Age phase, as this is the longest period of use of the necropolis. Even if the chronological indication for these two objects is not certain, their specific morphological features and the fact that both seem to belong to the most ancient contexts, it seems likely that they are more or less contemporary and represent the first appearances of these artefacts. Therefore, it can be assumed that arms of this type were widespread during this early time within funerary depositions while later their use shifted to temple, palatial, administrative, or domestic contexts. However, as there are no other examples of burial sites from this period in Cyprus or Anatolia, it is impossible to say anything with certainty.
Comparing the most numerous morphological groups (M3, M2, M4 and M1) with their geographical areas and the type of contexts in which they were found, it appears that only group M4 is found exclusively in the Anatolian area, and in six out of seven cases it belongs to a temple-type context. On the other hand, the other groups do not seem to be related either to a single type of context or to a single geographical area. Mielke (2006: 164-165) considers libation arms with a ware different from the RLWmW to be imitations. Manuelli (2009: 262-263) considers it more appropriate to speak of "different local productions" rather than of imitations 12 : the production of libation arms with different wares or with special surface treatments and a more accurate execution can be seen as a sign of Anatolian involvement in the creation and development of this form thanks to the contacts that took place over a long period in the area of the southern coast of Cilicia.
Because it was not possible to analyze the ware of all specimens, it is only possible to suggest possible morphological differences: the only two vessels that differ significantly from the standard are KU08 (Fig. 25) and BH01, where the right hand holds the cup from the side and not from the bottom. Rather than a local imitation or production, it has been suggested that this rendering reflects a misinterpretation or reworking of the most common type of libation arm. If these vessels were indeed local productions, they would still be an exclusively Central Anatolian product. However, as there is no more precise information on the type of ware or on the finding context 13 , this should be considered as a working hypothesis.
The analysis conducted so far, considering that the number of vessels showing any useful characteristics for this research is rather limited (80 individuals) and that it was based on morphological and macroscopic criteria, led to the conclusion that the existence of different production areas specialised in a specific morphological type can be considered another working hypothesis. Considering the Anatolian plateau, only three specimens belonging to specific morphological group were found outside North Central Anatolia, cf. appendix.
Recent studies show that a large group of RLWmW fragments are produced by the same fabric with a main workshop. From the analysis conducted on RLWmW samples, 9 arm shaped vessels from Kilise Tepe, Hattusa, and Tell Atchana were analysed. Those from Kilise Tepe and Hattusa have the same ware, and therefore the same workshop, as most of the pieces in RLWmW (Kibaroğlu et al. 2019: 416, 422-430 with further references). The number of ASVs analysed is only partially representative. Therefore, an analysis based on morphology is proposed as a working hypothesis. From a purely morphological and geographical distribution point of view, the proposal that more than one workshop may exist can be considered, as the chronological and contextual element is not discriminating (except for group M7). The M3 group, characterised by a more schematic and less accurate rendering of the hand with the fingers represented all at the same length along an imaginary line, seems to be widespread mostly in North-Central Anatolia (11 pieces) and covers a period ranging from the 15 th /14 th century BC to the 13 th /12 th century, contemporary to those from Cilicia (two pieces) and Cyprus (one piece). It is therefore possible to assume the existence of a Hittite production centre from which the pieces found in Cyprus and Cilicia were distributed. It should also be noted that this group is the only one found in Rough Cilicia, while the other geographical areas yielded evidence of several morphological groups, but the limited number of identified sherds makes it impossible to assume anything else.
Group M2, with its more naturalistic style, visible in the precise realisation of the nails and fingers in which the knuckles are sometimes also recognisable, shows vessels that were mainly disseminated in Anatolia (eight pieces) from the 14 th to the 13 th century BC. The only vessel from the Levantine area is dated to the same period. The vessels from Cyprus (four specimens) come from a funerary context, which is not represented in the other two areas, and are too broadly dated to allow a more precise determination. It is not possible to establish a single production centre for this morphological type, but it can be hypothesised the existence of a Cypriot production centre (with the oldest examples) and an Anatolian one (with the largest number of pieces). However, more data needs to be acquired in order to prove this hypothesis.
The M4 group, characterized by a more schematic and less accurate rendering of the hand but with the fingers converging in one point, seems to bring together fragments not only from Anatolia but also from the Upper Town of Boğazköy (six pieces), if we exclude the out-of-context vessels from Kaman-Kalehöyük. Therefore, it can be assumed the existence of a specialized centre for the production of arm shaped vessels on the site or in the surroundings of the Hittite capital as an example of a local production intended exclusively for temple use.
Group M1, with a more accurate rendering of the nails than M3, is attested only in Anatolian (six pieces) and Levantine areas (two pieces) from the 14 th to the 12 th century BC 14 . The hypothesis of an Anatolian production centre that exported this group in the northern Levant can be postulated.
The data collected on the geographical distribution of hand morphological types suggests the exclusion of a single centre of production of libation arms since it is not possible to define a single morphological type attested in a single period. Because there are several contemporary productions, it can be postulated that several workshops existed. However, the too broad dating of the contexts prevents the precise identification of these production cen-tres or the relationships between them, apart from the exceptions represented by groups M4 and M2. In any case, I think that the hypothesis of a single production centre covering the whole time span should be excluded.

FUNCTIONS
In the first analyses of these objects (Bittel 1937: 25-26), libation arms were associated with Egyptian censers. During the Middle Kingdom (ca. 2055-1650 BC; Bard [2013: 48]), an arm-shaped censer appeared in both visual representations and in the archaeological repertoire: these bronze censers, ended with a flat surface representing a hand (with no distinctions between right or left hand) with an open palm or, later, a papyrus plant. The opposite end was instead decorated with a hawk's head, which during the Middle Kingdom was turned outwards while in the New Kingdom it was turned inward (to the deity). A small container for storing the incense grains was often located on the arm, where, in later times, a figurine representing the kneeling pharaoh was sometimes added. The removable combustion chamber located on the hand was initially hemispherical and later of conical shape (Laisney 2009: 231-232) (Fig. 2). Since the inside of the arm is hollow, Laisney (2009: 248) assumed that the bronze covered a wooden core, to reduce the weight of the censer.
Egyptian pictorial representations (Ertem 1988: Fig. 31) also showed combustion chambers surmounted by small lines interpreted as flames or smoke from incense (Fig. 3). Given the similarity between Egyptian censers and the arms found in Anatolia and the Eastern Mediterranean, it was initially assumed that they had the same function (Bittel 1937: 25-26). However, if the Anatolian arms were used as incense burners, they would have shown a trace of combustion inside, either through fire or smoke. As there is no evidence of burning, they likely did not fulfil this function (Mielke 2007: 164).
By looking for a prototype to which the libation arms could have been inspired, Amiran (1962) identified comparisons in elephant tusks or bovine horns hollowed out on the inside with the smallest opening ending in the form of a very wide cup or spoon found in the Egyptian area. The specimens most reminiscent of ceramic libation arms were found in funerary contexts, which suggested a use related to deposition rituals. (Fig. 4). However, the association of the Anatolian and Eastern Mediterranean libation arms with the bovine horns suggested by Amiran as a possible prototype remains only an unverified assumption.
The most widespread opinion, given its shape and the fact that it was a hollow vessel, is that the function of the libation arms was to pour liquids during rituals. The contexts point towards the use of libation arms during rituals, as most of these containers were found in or near temple-type contexts. Following Eriksson's hypothesis (1993: 27), if they contained precious and perfumed oils, it can be suggested that the libation arms also had something to do with the use of these oils: perhaps oils were poured from the spindle bottles into the arm and from the arm onto a person or statue. By doing so, the bowl in the libation arm would receive the oil, that would be then mixed with water inside the arm before being poured out. The arm would therefore be used for anointing rather than libation (Mielke 2007: 164;Güterbock 1983).
However, also considering the lack of reference of such tools in more private contexts, with the exception of one specimen, it might be suggested that it was a personal object to be given to the deity as an offering, which might justify its presence also in the domestic sphere. An interesting change in function, yet unexplored, might have occurred from the end of the Middle Bronze Age, when the vessels was mainly recovered from funerary contexts to the 13 th century BC, when the arm shaped vessels come from templar contexts. Steel (2018: 204-206) does not doubt that they are still objects intended for libation but, in line with Mielke's analysis, sees them as containers intended for pouring liquid. Since there are no textual or iconographical information on how these objects were used, it can be assumed, given their bulky size, that these vessels required a certain amount of skill and experience to handle during libation. It has also been assumed, given the small capacity of the cup, that it was a container made especially for pouring that used an exact amount of liquid during libations. According to Harmanşah (2020: 235 with further references) the libation arm is associated with the Hittite term "GIŠ.ŠU.NAG.NAG or kattakurant" from Hittite ritual texts, which refers to a vessel in the shape of a cut or amputated arm widely used for libations of wine and other sacred liquids offered to the gods in sacred locations.
According to Pilavcı (2017: 245) the libatory function of these objects is undoubted but she redefines their function from containers that pour the liquid to containers that receive the liquid itself for the libation. In this case, the hand holding the cup represents the reception of the offering by the deity who, by presenting his outstretched arm, participates in the ritual. The interpretation of the vessels as the arm of the deity has been suggested for several reasons: the rendering of the hand in such an accurate and naturalistic way, compared to the schematic rendering of the part of the arm, might suggest that the tubular part was hidden under cloths that covered the statue, leaving only the part of the hand holding the cup visible. The vessel could be seen as an abbreviated form of the statue of the deity and thus placed on a surface or it could have been associated with a transformative value. Once poured, the liquid becomes immediately and directly accessible to the deity because it passes from the cup into the arm where it is contained. Furthermore, if the vessel is placed horizontally on a surface, it is possible that by filling it over the course of days the liquid was always present inside the cup, a fact that could be interpreted as the deity always being satisfied (Pilavci 2017: 221-225). Steel (2018: 204-206) wanted to shift the focus of the discussion regarding the RLWmW from the exclusive analysis of what was contained inside the libation arms to how these were used, noticing a substantial difference between the Cypriot and Anatolian contexts in which these artefacts were found. While the Cypriot contexts are mostly funerary contexts where the preservation of the pieces is good, the Anatolian contexts are mostly ritual and the libation arms found there are fragmentary or very poorly preserved. It is therefore clear that such objects in Anatolia were always available, in circulation, frequently used and just as frequently replaced. The differences between these two types of contexts show that there were different types of interaction between the objects and their users, which also reflect the different values attributed to them.
From the analysis of the artifacts, I find Pilavci's interpretation more likely. She sees these objects as something that receives the libation liquid during the ritual, rather than pours it. In fact, the size of a libation arm, combined with their weight that increases once the liquid is poured into them, makes the entire object difficult to handle and move. Moreover, the rim of the cup is often straight or not very everted, a condition that would make it very difficult to pour the liquid in a smooth and precise way. If these, as proposed by Pilavci, were placed on an inclined surface, the entrance of the liquid through the narrow passage that leads from the hand to the arm would be facilitated because the cup would never be filled. In this way everyone would be able to make such a gesture, even in domestic or private environments. I also agree with Steel's observations regarding the different type of interaction that occurs with the same type of object in Anatolia (fragmentary but abundant preservation) and Cyprus (more complete vessel but in funerary contexts): the almost daily use of the objects in Anatolia differs with their symbolic value and funerary function in the Cypriot contexts. This seems to be supported by the evidence from the contexts.

CONCLUSIONS
Based on these data, it is possible to postulate that the function of the libation arms is ritual. It is not yet clear, given the absence of mention of a similar object in written sources and figurative representations, whether it is a tool for rituals carried out by a priest, a private object to be used as an offering to a deity or used in private contexts as a representation of the deity himself to make an offering. In any case, any hypothesis that links them to incense burners or to any instrument that has to do with combustion is to be excluded, given the absence of traces of smoke or fire. It is more likely to be interpreted as an object into which the liquid is poured rather than one from which the liquid is poured.
The creation of a morphological typology demonstrates for the first time the existence of groups of libation arms with differences in the stylistic rendering of discriminating characteristics. By associating these groups with the contexts in which they were found and the regions to which they belonged, it is possible to detect that around the 16 th century BC in Anatolia the piece belonging to group M7 suggest a funerary function. There are no more examples of libation arms in funerary contexts after the 14 th century BC until we arrive at the almost exclusive association with temple-type contexts in the 13 th century BC both in Anatolia and in the Levant. The morphological typology also suggests the existence of two production centres during the Late Bronze Age: one in Anatolia (with groups M3, M4 and M1) and maybe another in Cyprus (most of the libation arms belonging to M2 were found in Anatolia but the vessels from the oldest contexts have been found in Cyprus. Also, the arm from Ugarit is more likely to come from a contact with Cyprus rather than with Anatolia).
Alongside the question of function is the question of the area of provenance. Cyprus can be reasonably excluded as the main place of production of these objects and this hypothesis can be confirmed by the fact that the Hittite world is well known to have included Hurrian and North Syrian religious practices in its culture (Mielke 2007: 164;Güterbock 1983). The written sources clearly state that the Hittites did not adapt any rituals from Alašiya (Cyprus) to their cult, and up to now there is no mention of such an object for libations in the numerous ritual texts found. Similarly, although the act of libation is often represented, there are no depictions of a similarly shaped vessel in Anatolia, even though most of the contexts in which they were found are templar-like. Furthermore, although the place of production of the entire ceramic class of RLWmW has been identified in Rough Cilicia, it should be noted that in the case of the libation arms this needs more evidence. Instead, these data seem to indicate Central Anatolia as the main place of production for this shape.  Fig. 1a