SURVEY, DRILL AND EXCAVATE. Complex geoarchaeological prospection of Bronze Age mounds as a key for understanding undermound architecture. A case study from Myluvannia, Western Ukraine

In the area of Western Ukraine, some aspects of mound (barrow) chronology might be resolved by using non‐ or minimally invasive archaeological prospection. As the cemeteries usually comprise two temporal units—the Late Neolithic Corded Ware Culture (third millennium BC) and Middle Bronze Age Komarów culture (second millennium BC)—by referring to particular funerary rites Komarów culture and their magnetic reflection, they can be distinguished by means of magnetometry survey supported with verification drilling. Recognition of the internal structure of the cemeteries is essential when confronted with their vast occurrence in Western Ukraine and also the morphological similarity of mounds. However, due to the great diversity in known grave architecture of the Komarów culture, it is essential to stress that the methodological approach presented here applies only to the specific construction type of burnt wooden and clay structures, which rarely appear in Corded Ware Culture funerary practices. In addition to excavation, the applied methods provided preliminary information on the cultural affiliation and time of the construction of two mounds in Myluvannia in the pre‐excavation stage of research. In this light, it is therefore possible to assess the potential occurrence of Komarów culture mounds in other Western Ukrainian mound cemeteries.


| INTRODUCTION
For over 10 years, there has been an intensification of research on mound (barrow) cemeteries in Western Ukraine by the joint Polish-Ukrainian team from Pozna n and Kyiv (Makarowicz et al., 2013;Makarowicz et al., 2018Makarowicz et al., , 2019Makarowicz et al., 2021;Makarowicz, Kochkin, et al., 2016;. This area was characterized by cultural complexes during different prehistoric periods. Subsequent research campaigns as well as archival information confirm this cultural and chronological diversity, as well as the various ritual behaviours and patterns used by the prehistoric communities who built these mound necropolises.
As a result, this research provides a framework for introducing and testing a whole set of multidisciplinary methods. In this article, we describe the results of previous research with novel techniques to identify particular types of funeral features in the exploratory stage prior to archaeological excavation. The main goal is to define a specific type of funeral architecture that can be identified during noninvasive research in Western Ukraine.

| Mound archaeology and magnetometry
Research on prehistoric round earthen mounds (henceforth mounds) comprises various methodological approaches. Regardless of chronology, several aspects of mound construction are currently being brought to attention and span both macro-and microscale studies.
The former attempt to investigate the spatial distribution of the mounds via GIS analyses to evaluate their position in the landscape (e.g. Cwali nski & Niebieszcza nski, 2012;DeReu et al., 2011;Eckardt et al., 2009), arrangement patterns and alignments to celestial bodies (e.g. Arnoldussen & Drenth, 2015;Makarowicz et al., 2019) or in some cases, to trace their nonfuneral function as possible orientation landmarks along communication routes (Holst et al., 2001;Johansen et al., 2004). On the other hand, the micro-approach consists of numerous aspects that include multidisciplinary methods derived from various branches of science like geology, chemistry and biology. In addition to the main purpose of revealing the chronology and funerary aspects of the mound builders, this type of research can be oriented towards palaeopedological (e.g. Doorenbosch & van Mourik, 2016;Holst et al., 2001), palaeoclimatological (e.g. Hildebrandt-Radke et al., 2019) and micromorphological aspects (e.g. Cammas, 2018;Macphail et al., 2013), as well as other applicable 'archaeological sciences' depending on the research aim. However, the main tool for obtaining the necessary data (i.e. soil samples, bone material, ceramics and charcoal) is still excavation-which in the light of existing analytical options has evolved into highly detailed and precise fieldwork. This implies the use of vast amounts of funds and labour to gather quality data and as such demands that the cultural and chronological affiliation of mounds must be recognized pre-excavation to minimize potential risks of conducting fieldwork on inaccurate assumptions. Current archaeology uses several noninvasive methods represented mostly by geophysical prospection. The application of geophysical techniques to investigate past human activities is considerable enough to be recognized as a separate branch of science known as archaeo-geophysics (El-Qady & Metwaly, 2019;Malouchos et al., 2021;Sala et al., 2012;Schurr et al., 2020). Three geophysical methods-electrical resistivity of strata, ground-penetrating radar (GPR) and magnetometry-are commonly used in archaeological prospection to provide a glimpse into the vertical stratigraphy of archaeological deposits as well as their spatial distribution. Concerning mound archaeology, all three are applicable; however, their utility is restricted by several factors like the construction material used in the undermound architecture (i.e. stone constructions and daub material) or relative mound height (e.g. Hildebrandt-Radke et al., 2018;Wroniecki et al., 2017). The latter often excludes usage of magnetometry as a preexcavation tool due to the risk that deeply buried funeral features will not be registered (Makarowicz et al., 2017) or, more probably, will be misinterpreted. However, in the case of shallower mounds, the application of magnetometry can result in a two-dimensional plan indicating the position of anomalies that reflect the presence of various features.
In combination with archaeological experience and contextual knowledge, the interpretation of magnetic imagery is often a useful tool for preliminary chronological and cultural estimations because it can distinguish particular and specific features (e.g. stone rings, stone kernels; mortuary houses-see, e.g., Hildebrandt-Radke et al., 2018;Makarowicz, 2019;. To prevent misinterpretation and, consequently, costly excavations of unexpected archaeological contexts, additional methods are now being applied, including soil sample analyses and other geophysical techniques like magnetometry or GPR. The use of geophysical methods on each mound is especially necessary in the case of multicultural cemeteries (i.e. the inclusion of later period mounds in Late Neolithic and Bronze Age mound alignments).
This study provides an example of a magnetometry survey of a Middle Bronze Age mound necropolis in Western Ukraine with additional drilling verification of two mounds, AMS 14 C dating of relevant organic material, and further excavations of selected mounds to extrapolate the results onto the remaining burials within the cemetery, thus providing a more accurate interpretation of mound chronology and cultural affiliation.

| Middle Bronze Age mounds in Western Ukraine
During the Middle Bronze Age, a vast area of Central and Central-Eastern Europe became the scene where societies who erected mound cemeteries constituted a special burial practice related to cultural change. In the eastern portion of Central Europe, Trzciniec Cultural Circle societies also practiced this burial custom (Makarowicz, 2010, pp. 201-280;pp. 369-386), as is evidenced in its south-eastern branch represented by the Komar ow culture Romaniszyn, 2015;Romaniszyn et al., 2017;Sulimirski, 1968;Swiesznikow, 1967). However, in the case of Western Ukraine, the pioneering time of mound occurrence, as earlier Late Neolithic Corded Ware culture societies (2800-2200 BC), also buried their deceased under mounds (Czopek et al., 2015;Machnik et al., 2006aMachnik et al., , 2006bMachnik et al., , 2006cMachnik et al., , 2011Sulimirski, 1968). The Komar ow culture societies (1800-1500 BC) often raised their burial mounds within pre-existing Corded Ware culture alignments, thus creating multiphase necropolises.
In addition, some Late Neolithic and Middle Bronze Age monuments were reused in later periods (i.e. Late Bronze and Early Iron Ages- Dascalu, 2007;Dergachev & Sava, 1984;Roman Period-Smishko, 1960;Vakulenko, 2008). The cultural landscape of the area is therefore a mix of different successive cultures represented by the same morphological remnants of their burial customs-mounds. In light of archaeological data, the Corded Ware culture and Komar ow culture mounds are often similar in size and on-surface appearance (Makarowicz, Kochkin, et al., 2016), but they differ considerably in terms of undermound architecture. While the former usually comprise singular skeletal burials with associated funerary objects (Włodarczak, 2006(Włodarczak, , 2017 and sometimes a palisade made of posts designating the perimeter of the mound (e.g. Czebreszuk & Szmyt, 2011), the latter often served not only for burials but also included additional mortuary structures (Makarowicz, 2019;Romaniszyn, 2015). The Komar ow culture people created a complex of underground architecture, among which the so-called cenotaphs   Often, additional deposits of cremated human remains were included within the mound during its erection. The Corded Ware culture mounds in this area, on the other hand, usually did not comprise such advanced architecture, as the mound was raised directly over the grave.
In terms of archaeological prospection, the different chronological stages of cemetery development, which are complicated further by diversified burial practices within the same cultural unit, create on the one hand a significant problem when interpreting the results but on the other serve as a testing field for assessing the use of magnetometry and additional verification methods and possibly extrapolating these interpretations. Such a case is present at the Myluvannia necropolis-a part of a vast Bukivna cemetery complex-where several mounds were prospected using magnetometry and drilling followed by excavation, thus providing the opportunity to compare the underground architecture with its appearance on the magnetometry plans.

| Study area and geological parameters
The Myluvannia cemetery, a part of the Bukivna cemetery complex, is mounds occur is overgrown by old broadleaf forests, thus preserving the mounds and preventing their destruction due to agriculture (Makarowicz, Kochkin, et al., 2016, p. 39).
The surficial geology of the area is represented mostly by loess of differentiated thickness deposited since the Pleistocene (Makarowicz, Kochkin, et al., 2016, pp. 37-38). The upland is intersected by numerous deep incisions from riverine erosion and gullies ( Figure 1).
The appearance of loess deposits provided the opportunity for chernozem to develop in the past, which was documented under several burial mounds in Bukivna as a palaeosol level

| Cemetery in Myluvannia
The necropolis in Myluvannia is the eastern extension of the vast cemetery complex of Bukivna (Figure 1). The latter comprises a total of 149 mounds spread in several groupings of linear and group-like arrangements (see Makarowicz et al., 2019;Makarowicz, Kochkin, et al., 2016). The western part of the complex, located near Bukivna village, was previously subjected to excavation (Bryk, 1932;Siwk owna, 1938;Makarowicz et al., 2013;Makarowicz, Kochkin, et al., 2016, pp. 184-212). A recent excavation project revealed the presence of Late Neolithic mounds between which Middle Bronze Age Komar ow culture mounds were inserted (Makarowicz et al., 2018). The eastern part of the cemetery located north-east of Myluvannia village numbers 56 mounds which are arranged into three groupings. Our study was oriented towards the central group

| Research methods
This study utilized several methods to reveal the chronological and cultural affiliation of Myluvannia Group 1 (Figure 3). It first combined forest reconnaissance and noninvasive prospection along with geoarchaeological drilling with AMS 14 C-dating of the resulting organic material (i.e. charcoal). The results were verified through excavations of two selected mounds (259 and 260) and magnetic susceptibility.

| Forest reconnaissance and topographic survey
The site was discovered in early spring of 2016 during a forest survey of a woodland area near Myluvannia village. A group of archaeologists with hand-held GPS receivers prospected the area, which resulted in the location of three mound groups. Each was then surveyed by means of Leica TCR laser tachimeters to construct their digital elevation models. These models were georeferenced into the Pulkovo 1942 coordinate system of Zone 5.

| Geophysical prospection
After geodetical documentation, Group 1 was prospected using the noninvasive geophysical method of magnetometry. A total of 10 mounds forming the main line were investigated as were the spaces between the mounds, covering an area of 0.76 ha (Makarowicz, Kochkin, et al., 2016). The survey polygon was divided into a net of grids, each measuring 20 Â 20 m and fixed in the

| Excavations of Mounds 259 and 260
Excavations conducted on the two selected mounds evidenced the presence of the remains of complex Komar ow culture funerary rites and undermound architecture ( Figure 5). In Mound 259, three specific features were distinguished. The first (259/F1) was the remnants of a burnt wooden plank which indicate both the existence of a wooden structure as well as its burning. This feature included a deposit of three ceramic vessels belonging to the Komar ow culture and a concentration of human bones directly to the north. Wooden constructions are a common element of the funeral custom within the complex rites of the Komar ow culture (Romaniszyn, 2015). Another feature (259/F2) was an agglomeration of small stones (5-10 cm) positioned along the eastern margin of the mound. The most intriguing feature revealed within the mound was a double-horse burial, oriented with the heads facing each other, which finds its analogy among known Komar ow Culture funerary practices at, for example, Bukivna or Husiaty n (Bryk, 1932;Ilchyshyn, 2016; see Przybyła, 2020).
However, a typical funerary feature of the Komar ow Culture was revealed in Mound 260. In its north-eastern portion, a large ritual object was documented and interpreted as a mortuary house (260/F1). This construction was built of wood and clay, inside of which the deceased was lain. After the burial, the whole structure appears to have been ritually burned, and as a result, the clay was fired into daub. In the end, the entire area of the ritual was covered by earth (blocks of turf), thus creating the mound. In the case of Mound 260, it is more than evident that this is a mortuary house. The documented profile and flat plan show that the floor was constructed from clay while the walls were raised using wooden planks additionally enforced with clay material ( Figure 5). However, the entire feature was not preserved in its entirety-the plan of the mortuary house was reconstructed based on the distribution of daub, charcoal and bone fragments as well as vessel deposit locations, which approximated a rectangular shape ( Figure 5). Feature 1 was oriented along the SW-NE axis and might have been partially open towards the latter.
The stratigraphy of Mound 260 revealed during the excavations corresponded with the drilling results.
F I G U R E 4 Magnetometry results of the Myluvannia cemetery's Group 1 (a) (Makarowicz, Kochkin, et al., 2016) and their close up view on Mounds 259 (c) and 260 (b). Legend: 1: mound outlines based on the occurrence of magnetic anomalies related to archaeological features; 2: course of magnetic and morphological profiles (see Figure 6); 3: outline of central anomaly within the barrow 260 and related probably to the burnt archaeological feature

| Magnetic susceptibility
To find whether a large feature corresponded with the magnetometry prospection results, three samples from Feature 260/F1 were subjected to magnetic susceptibility measurements (Table 1)  with fired clay material, it should be considered as the feature that produced the anomaly. Moreover, this interpretation is supported by the measurements of magnetic susceptibility, which indicate the presence of significant volumetric susceptibility (Table 1). Because it was burned, the feature is clearly discernible on the magnetic plan even though it was covered by a mound. Also, the distribution of the magnetic anomaly along the transect ( Figure 6) overlaps with the extent of the mortuary house seen in the excavation profile. The same is evident when the anomaly is compared with the distribution of charcoals and daub fragments discovered within the feature.
On the other hand, the anomaly registered in the centre of Mound 259 was not supported by the occurrence of a similar feature to that in Mound 260 ( Figure 6). Nor can an archaeological feature registered within excavation trenches be attributed to its position.
Therefore, it is highly probable that the bipolar anomaly was caused by the presence of a modern object directly within the forest undergrowth. This assumption might be supported further by the presence of a slight depression observable at the summit of the mound-possibly traces of an old looters' trench. Nevertheless, various archaeological features occurred within or under the mound.
It is therefore essential to stress the lack of their clear magnetometric manifestation. The main explanation for this might lie in the nature of these features. As none were burned, they are harder to be identified using magnetometry. Additionally, the double-horse burial in Mound 259 also lacked any magnetic manifestation, because the feature was submerged beneath the original surface level (ancient humus) and probably due to the occurrence of strong anomaly on top of the mound, not related to any archaeological feature.  (Włodarczak, 2006(Włodarczak, , 2017 (Romaniszyn, 2015;Sulimirski, 1968). Therefore, we underline that the proposed methodological key concerns only this specific burial feature.
Based on the similarity of the magnetic appearance of the undermound constructions made of wood and clay and further altered by fire, it is possible to distinguish two groups of mounds in Myluvannia cemetery Group 1. The first consists of the verified Mound 260 as a reference and Mound 276, which is characterized by the occurrence of the same magnetic anomaly in its centre. These anomalies are identical in terms of their extent, magnetic field arrangement, orientation along the NNE-SSW axis and recorded nT values (from 16 to À8 nT/m). It is highly probable that Mound 276 was constructed in the same pattern, consisting of a mortuary house made of burnt organic and clay material. This, however, must be verified through at least drilling, as a slightly invasive method, which we pro-

| CONCLUSIONS
The complexity of the funeral rites observed under mounds in Western Ukraine offers the optimal testing ground for the identification of grave structures using noninvasive methods. The results of magnetometry prospection carried out in recent years clearly revealed the presence of complex funeral structures beneath the mounds.
Features that were burned during various rituals attract special attention. These were mortuary houses in which the deceased and grave goods were burned. The verification of the magnetometry prospection results with the analysis of drilled cores allowed these structures to be identified. This aspect can be helpful for the chronological or cultural recognition of a mound. The burning of grave features in the mounds of the upland, loess zone was quite a common pattern for communities from various cultures and periods (not only those of the Middle Bronze Age but also those during the Iron Age and Roman Period). We are aware that such estimations are highly arguable without excavation data, but we do believe that this methodology will aide in the selection of appropriate Middle Bronze Age Komar ow culture mounds for excavation.