Neolithic Fishing Stations at Šventoji, Southeastern Baltic

Abstract

Several examples of de-Neolithisation have been identified in the Eastern Baltic region, where communities of Neolithic cultures—particularly in areas rich in aquatic resources—shifted from animal husbandry to fishing, or at least significantly supplemented their subsistence with freshwater resources. One such case is Šventoji in Lithuania, on the southeastern Baltic coast, primarily known for its numerous waterlogged sites with excellent preservation of wooden artefacts, dated to the Subneolithic (ca. 4000–2900 cal BC). In 2021 and 2024, investigations of three short-term stationary fishing sites, dated to the earlier part of the Neolithic (ca. 2850–2500 cal BC), provided an opportunity to compare Neolithic stationary fishing with its Subneolithic counterparts. At the beginning of the Neolithic, we identified a technological shift in the construction of fish weirs and traps: pine laths were replaced by round shoots of deciduous trees. In addition, exceptionally rare archaeological finds were uncovered at the Neolithic fishing sites of Šventoji—two stone battle axes with preserved wooden handles. Accordingly, alongside the study of Neolithic stationary fishing, this paper presents the results of radiocarbon dating (¹⁴C), taxonomic identification of the wood, and use–wear analysis of these two axes.

1. Introduction

Genetic studies indicate that in most of Europe, farming and animal husbandry spread primarily through the migration of Neolithic populations [1,2,3,4]. Expanding Neolithic farmers and herders brought new genes, new ways of life, and new diets into hunter–gatherer–fisher territories, and transformed the landscape by clearing forests and converting it into pastures and arable fields [5,6,7,8,9]. Nevertheless, the timing, scale, and pace of these changes varied considerably across the continent. In Northern Europe, where the climate was colder and less suitable for agriculture, indigenous populations were not assimilated for centuries: they lived alongside Neolithic farmers while continuing to hunt and fish. For instance, the Pitted Ware culture in Sweden, which preserved a Mesolithic lifestyle and genome, disappeared only around 2200 cal BC [10], whereas in northeastern Lithuania, local populations maintained Mesolithic traditions until the very beginning of the Bronze Age, ca. 1800 cal BC [11].

Moreover, on the periphery of the Neolithic world, the archaeological record provides evidence for a counter-process to the expansion of agriculture—its decline, or the so-called de-Neolithisation. For instance, in southern Norway, the agricultural package appeared around 4000 cal BC, yet within only a few centuries the earliest farming communities reverted to the exploitation of marine resources [12,13]. Certain indications of de-Neolithisation are also evident in eastern Denmark ca. 3000 cal BC [14]. Few cases of de-Neolithisation are known from the Eastern Baltic, where communities of Neolithic cultures—especially in areas rich in aquatic resources —abandoned farming and herding in favour of fishing, or at least substantially supplemented their diets with freshwater resources [15,16]. One such region is Šventoji, Lithuania, located on the southeastern Baltic coast. During the Subneolithic and the earlier part of the Neolithic (ca. 4000–2400 cal BC), the area was occupied by a shallow, fish-rich lagoonal lake, on the shores and within the basin of which approximately 50 dwelling and fishing sites are currently known (Figure 1) [17]. Most of these belong to the Subneolithic pottery-making hunter–gatherer–fisher communities. Thanks to over 50 years of long-term and extensive archaeological excavations covering approximately 3000 m², Subneolithic fishing practices in the Šventoji lagoonal lake are now relatively well understood and have been the subject of several scientific publications [18,19,20].

Around 2900–2800 cal BC, Neolithic groups appeared in the Šventoji region—specifically, the Globular Amphora (GAC) and Corded Ware (CWC) cultures, both of which were primarily herding societies. Evidence that they also engaged in fishing in the Šventoji lagoonal lake is provided by their pottery, recovered from lakebed contexts at Šventoji 1, 2/4, and 6 sites [17,18,20]. Until recently, however, evidence for short-term, isolated Neolithic fishing sites was lacking. Without such sites, it was impossible to determine whether, and how, fishing traditions in the lagoonal lake had changed at the very onset of the Neolithic—that is, with the arrival of these new groups. In 2021 and 2024, such an opportunity finally arose. In the northern part of the Šventoji archaeological complex, three short-term stationary fishing sites, dated to the earlier part of the Neolithic (ca. 2850–2500 cal BC), were identified and partially investigated.

The aim of this article is to examine stationary fishing in the Šventoji palaeolake based on these sites, to identify changes at the beginning of the Neolithic, and to establish their chronological framework. In addition, we present two stone battle axes with preserved wooden shafts, one of which was recovered during our excavations in a Neolithic fishing site. These are the only known Corded Ware battle axes with wooden handles preserved, and both were discovered at Šventoji fishing sites.

2. Site Description

The archaeological complex of Šventoji in northwestern Lithuania was discovered during drainage works at the end of the 1960s (Figure 1B). At the same time, archaeological excavations began [18] and have continued to the present day [17], albeit with occasional interruptions. Today, the complex comprises over 60 sites dating from the Late Mesolithic to the Late Bronze Age (ca. 6000–500 cal BC) (Figure 1A). It is particularly known for the well-preserved artefacts made of plant and animal materials, excavated at Subneolithic and Neolithic wetland sites dating to ca. 3500–2500 cal BC [18,21,22]. Many of these sites are associated with fishing activities and are located in the bed of a former lake that originally formed in the place of a lagoon [20]. Unfortunately, due to the lowering of the groundwater table during extensive land reclamation in the 1960s–1970s, the wood artefacts at many previously rich sites (e.g., 6, 23, and 26) have severely decayed or completely disintegrated. Moreover, the Šventoji archaeological complex is currently threatened by rapid urbanisation associated with the expansion of the nearby town of Palanga [17].

In this article, we present new data from six Subneolithic and Neolithic sites at Šventoji: 1, 42, 41A-1, 41A-2, 41A-3, and 41A-4. Šventoji 1, the richest site in the region, was excavated by R. Rimantienė between 1967 and 1969. During three field seasons, an area of 1860 m² was investigated, yielding an exceptionally rich assemblage of Subneolithic Porous Ware and Neolithic CWC ceramics recovered from the gyttja of the former lake [18]. It appears that a nearby dwelling site was situated on a sandy island that had been ploughed for many years, leaving only amber and flint artefacts preserved there. R. Rimantienė noted that in some areas of the littoral zone, artefacts of different chronologies were found within distinct gyttja horizons. At both levels, in addition to pottery, amber, and stone finds, artefacts made of plant and animal materials were also well preserved. These included sections of fish fences made of pine laths, wooden leister prongs, fragments of troughs and logboats, handles and sleeves of stone axes, fish-driving sticks, paddles, net poles, and other items. Radiocarbon dates indicate that the main phase of activity at Šventoji 1 occurred between 3200 and 2500 cal BC [18]. In 2022 and 2024, G.P. surveyed the site through systematic test-pitting (301 m²) and excavated three trenches (81 m²), confirming that the Subneolithic fish fences and the archaeological layer—locally enriched with freshwater fish bones—are still preserved (unpublished data). For the present paper, we analysed only a single artefact from R. Rimantienė’s excavations—a stone battle axe with its wooden handle preserved.

Šventoji 42 was discovered by G.P. during the supervision of sewage pipe construction in 2009. The mechanically excavated trench measured 934 m in length, 1.5–8 m in width, and 2–5 m in depth, allowing for the documentation of a detailed sedimentation sequence within a profile that cuts across almost the entire area of the former lagoon [23]. In the middle and deepest part of the former lake, the remains of wooden constructions were documented, extending for approximately 50 m. In 2011, an additional trench measuring 1 × 43 m was excavated at the site. Across both trenches, 14 wooden piles measuring 4–8 cm in diameter were recorded, along with numerous fragments of pine laths between them. In addition, four stone net sinkers and one pine-bark net float were found. Based on the limited number of finds, it appears that the site was used briefly for fishing, employing both nets and fish fences and/or traps made of pine laths. Two wooden piles were dated to ca. 2900–2750 cal BC, corresponding to the very end of the Subneolithic period [23]. In this article, we report three additional dates from wooden piles.

The name Šventoji 41 refers to an area approximately 200 × 200 m in size, located at the northernmost edge of the archaeological complex (Figure 1A). In 2008 and 2012, the area was surveyed by G.P. through test pits (Figure 1C), revealing artefacts from different periods, each occurring within distinct stratigraphic units. Subneolithic and Neolithic finds, mostly associated with fishing activities, were recovered from the lower and middle parts of the lacustrine gyttja, respectively, and were named Šventoji 41A. Meanwhile, Bronze Age material (ca. 1500–800 cal BC) was found within a later riverbed that had cut through the lacustrine sediments and was named Šventoji 41B [17]. In 2021 and 2024, more detailed excavations were carried out. G.P. investigated seven trenches, each 3 m wide and 15–27 m long, in the western part of the area and identified the Neolithic fishing site 41A-1. In 2024, A.K. and K.P. mechanically opened two large plots, measuring 542 and 152 m², and identified the Neolithic fishing sites 41A-2 and 41A-3, as well as one Subneolithic site, 41A-4 (Figure 1C). Unfortunately, none of the fishing sites discovered, perhaps with the exception of 41A-3, were fully excavated, as the landowners financed work only in the areas of planned construction.

The Šventoji 41A-1 fishing site covers an area of at least 12 × 8 m, although only its eastern part, measuring approximately 6 × 8 m, was excavated (Figure 2A). The remains of the site were found in gyttja at a depth of approximately 0.8 m. The gyttja was stratigraphically heterogeneous, consisting of two layers of roughly equal thickness (about 55 cm each). The upper layer was brown and cracked due to drying, whereas the lower layer was olive-grey, wetter, and sandier in texture. In some locations, the two gyttja layers were separated by thin (1–2 cm) lenses of small plant remains.

Eight wooden piles, 5–8 cm thick, were found in the trenches. They had been driven through the lake-bottom mud and up to 30 cm into the underlying marine sand. However, a significantly larger number—25 thinner stakes, 1–2 cm in diameter—had been inserted into the lower part of the gyttja, without their tips reaching the marine sand (Figure 2). In addition to the piles and stakes, two concentrations of horizontally lying, heavily decayed shoots, 0.5–2 cm thick, were recorded, each approximately 1.5–2 m long and 0.5 m wide. The shoots lay in a single direction, arranged in two or three layers side by side (Figure 2). A fragment of a wooden artefact, possibly a paddle blade, was found between the concentrations of shoots (Figure 3(1)). Five stones were discovered at the fishing site, but only two of them lay within the Neolithic horizon; the others occurred at deeper levels. These were unworked, naturally rounded sandstone and granite beach pebbles, measuring 17 × 13 × 3 cm and 6 × 4 × 2 cm, respectively (Figure 3(2,3)).

Fishing sites 41A-2 and 41A-3 were discovered approximately 150 m from 41A-1, within an excavated area of 542 m², 45 m in length and about 12 m in width (Figure 1C). At 41A-2, beneath a 0.4 m thick ploughed layer, there was a ca. 1.2 m thick gyttja deposit—black at the top, grey-brown in the middle, and olive-grey at the bottom. Below the lacustrine sediments lay either fine sand of the Littorina Sea or a buried Mesolithic soil consisting of peat and/or peaty sand (Figure 4B). In the lower part of the gyttja and in the marine sand, six wooden piles measuring 4.3–6 cm in diameter were found, spaced 0.5–4 m apart (Figure 4A). At a distance of 3–6 m from the piles, at a depth of 1.4 m in a brown gyttja layer, two concentrations of wooden shoots 1–1.5 cm thick were uncovered (Figure 4A,C). The western concentration occupied an area of approximately 1.8 × 0.7 m, while the eastern concentration was elongated, tapering at one end, and measured 1.6 × 0.7 m. The shoots lay parallel, in two layers, with no traces of binding observed. Only one stone artefact was found in the Neolithic horizon: a flaked quartz sandstone plate measuring 11 × 7 × 2 cm (Figure 3(4)).

The 41A-3 fishing site was located 15–25 m southeast of 41A-2. Here, a similar stratigraphy was documented, except that the gyttja layer was thinner—0.6–1 m (Figure 5B). A group of nine vertical piles extended over a ca. 12 m strip, spaced 0.5–3 m apart (Figure 5A), with diameters ranging from 4 to 5.5 cm. Approximately 3.5 m west of the piles, at the contact horizon between the brown and olive-grey gyttja and at a depth of 1.2 m, a stone battle axe with a preserved wooden handle was discovered (Figure 6). The blade faced upwards, while the butt end had sunk slightly deeper (−0.11 to −0.16 m a.s.l.). At a distance of 2.5 m from the axe, at a similar depth, a notched quartz sandstone slab measuring 23 × 17.5 × 3 cm was found, bearing two flake scars on opposite sides (Figure 3(5)).

Šventoji 41A-4 is a Subneolithic fishing site, the remains of which were uncovered during the mechanical excavation of a 152 m² area (Figure 1C). In the lower part of the gyttja layer, within an area of approximately 5 × 3 m, two alder piles (6 and 8.5 cm in diameter) and four pile holes were recorded, while numerous fragments of pine laths were dispersed over a wider area.

3. Materials and Methods

In Šventoji, an area of about 63 ha is protected by the state as cultural heritage. Prior to any construction, trial archaeological investigations are required in the area. The methods for these investigations are defined in the Heritage Maintenance Regulation, “Archaeological Heritage Maintenance” [24]. Unfortunately, this document does not take into consideration the unique characteristics of the archaeological site. According to the Regulation, during trial investigations, at least 1% of a surveyed area must be tested by test pits. This method is suitable only for the detection of comparatively large-area archaeological settlements. However, nearly 20 years of research in Šventoji have shown that test pitting is ineffective for locating Stone Age fishing sites with sparsely placed wooden stakes and few other archaeological finds. When small test pits are dug every 10–20 m, the remains of such fishing sites typically remain undetected. For example, Neolithic sites 41A-2 and 41A-3, as well as site 41A-4 of a slightly earlier period (Subneolithic), remained undetected by test pitting in 2008 (Figure 1C). Anticipating this limitation, in 2024 we carried out broader excavations even in areas where no archaeological finds had been recovered from the test pits. During these investigations, peat and lacustrine deposits were removed with an excavator, opening trenches of varying lengths and 3–5 m in width, which were later joined to form larger excavation areas (Figure 1C). Manual excavation was employed only when archaeological wood was encountered, allowing for the careful exposure and documentation of the fishing site remains.

The anatomical–microscopic analysis of wood was carried out by K.P. Wood samples, approximately 1–2 cm in size, were collected from the archaeological wooden finds for thin-section preparation. Tree species were identified for all piles and stakes. However, only 3–8 samples of the numerous shoots from each accumulation were examined, as the wood was poorly preserved. Only one artefact had been conserved—the handle of the stone axe from Šventoji 1. Polyethylene glycol (PEG) was removed prior to analysis by soaking the sample in warm water. Thin slices in transversal, radial, and tangential orientations were produced by hand-cutting the samples with a razor blade and mounting them on microscope slides. A few drops of glycerol were applied to the slices, and cover slips were placed on top to secure and preserve the specimens for microscopic analysis. The species identification was performed using a Zeiss Axioscope 5 microscope, with thin sections of wood analysed in transverse, tangential, and radial planes. The assessment of microscopic wood features was based on wood anatomy atlases [25,26]. In some cases, the outermost annual rings were preserved intact, as indicated by the presence of bark or well-preserved phloem. In these instances, the season of tree death (felling) was determined by evaluating the anatomical structure of the last ring [27,28]. The following stages of last ring formation were distinguished: Stage A—early part of the last ring forming: the tree was felled at the beginning of the vegetation season; Stage B—early part of the last ring fully formed, late part not yet formed or still forming: the tree was felled in the middle or second half of the vegetation season; Stage C—the last ring fully formed: the tree was felled during the non-vegetation season. When it was not possible to reliably determine whether the last annual ring was still forming or already fully formed, intermediate stages A-B and B-C were introduced.

Thirteen wooden samples were radiocarbon dated in this study. Eleven were obtained from wooden constructions and two from stone axe handles. The outer rings of piles and stakes were sampled with a knife, taking care to avoid any modern roots growing into the old wood. Since all dated objects were made from young trees of short-lived species, the old wood effect can be considered negligible. Nearly all samples were taken immediately after excavation and required no contamination screening or additional cleaning. Only a single sample was taken from an already conserved artefact—the wooden handle of the Šventoji 1 stone axe. The sample was soaked in acetone, alcohol, and hot water to remove polyethylene glycol (PEG). It was subsequently screened for any remaining contamination using FTIR.

Radiocarbon dating was conducted at three laboratories using different equipment, as it was organised out by different archaeologists during three campaigns shortly after the excavations in 2011, 2012, and 2024. In 2011, the ¹⁴C content of three wooden samples was measured by liquid scintillation counting (LSC) at the Laboratory of the Institute of Geology, Tallinn University of Technology (Estonia). Later, AMS ¹⁴C dating was carried out in two laboratories: the Centre for Physical Sciences and Technology in Vilnius (Lithuania) and the Poznań Radiocarbon Laboratory (Poland). In all three laboratories, the standard acid–alkali–acid (AAA) pre-treatment was applied to the wood samples. Detailed descriptions of the sample preparation and quality control methods employed at the Vilnius AMS dating lab have been published recently [29]. The Poznań laboratory implements stringent quality control measures, including: (1) chemical pretreatment protocols to remove contaminants; (2) precision checks on modern standards and background samples; (3) participation in interlaboratory comparison programmes such as VIRI and SIRI; and (4) comparison of ¹⁴C measurements between two AMS spectrometers to ensure measurement reliability [30]. In this study, all radiocarbon dates were calibrated using OxCal 4.4 software and the IntCal20 atmospheric curve [31,32]. Calibrated dates are presented at 95.4% probability.

Microscopic analysis of technological and use–wear traces was carried out by G.O. on two artefacts—stone battle axes from the Šventoji 1 and 41A-3 fishing sites. The analyses were performed using a Nikon SMZ-745T microscope (up to 65× magnification) equipped with a DeltaPix Invenio 6EIII camera, which was used to obtain the photomicrographs. For logistical reasons, it was not possible to conduct the use–wear analysis with a metallographic microscope. In the case of the axe from 41A-3, only the handle was subjected to traceological analysis; the axe itself was not examined due to heavily contaminated surfaces with lake sediments and post-depositional alterations. The axe from the Šventoji 1 site, however, was analysed in its entirety. Traceological research on wooden artefacts has only recently begun, and a methodology that fully accounts for their specific characteristics has not yet been developed. Consequently, the reported microscopic studies are based entirely on the methodology and terminology developed for traceological studies of stone and bone specimens (e.g., [33,34,35,36,37,38]).

4. Results

4.1. Fish Fences and Traps

At Šventoji 41A-1, piles with diameters of 5–8 cm were sharpened either by splitting off long flakes around the entire perimeter (Figure 7(5,6)), only on one side (Figure 7(3)), or by chopping out two small opposing flat surfaces (Figure 7(4)). The thinner stakes were only slightly shaped—by removing a single long flake (Figure 7(1,2))—and sometimes they were simply pushed into the sediment without sharpening. The piles were made of alder (7/8) and ash (1/8). The thin stakes were made from hazel (18/25), species of the family Salicaceae (6/25), and species of the family Rosaceae (1/25) (

Table 1. Wood species and stages of final annual ring formation in wooden artefacts from Neolithic fishing sites Šventoji 41A-1, 41A-2, and 41A-3.

	Horizontal Shoots (0.5–2 cm)			Vertical Piles (4–8 cm)						Vertical Stakes (1–2 cm)					Total
Site and Species	B-C	N/I	Total	A	B	B-C	C	N/I	Total	A-B	B-C	C	N/I	Total
41A-1	4	5	9	2	–	9	–	5	16	1	9	3	3	16	41
Alnus sp.	–	–	–	–	–	4	–	3	7	–	–	–	–	–	7
Corylus avellana	2	2	4	2	–	4	–	–	6	1	6	2	2	11	21
Fraxinus excelsior	–	2	2	–	–	–	–	1	1	–	–	–	–	–	3
Rosaceae	–	–	–	–	–	1	–	–	1	–	–	–	–	–	1
Salix sp.	2	1	3	–	–	–	–	1	1	–	3	1	1	5	9
41A-2	–	7	7	–	1	–	1	4	6	–	–	–	–	–	13
Alnus sp.	–	3	3	–	–	–	–	1	1	–	–	–	–	–	4
Corylus avellana	–	4	4	–	–	–	1	2	3	–	–	–	–	–	7
Fraxinus excelsior	–	–	–	–	1	–	–	1	2	–	–	–	–	–	2
41A-3	–	–	–	1	–	–	–	8	9	–	–	–	–	–	9
Alnus sp.	–	–	–	–	–	–	–	5	5	–	–	–	–	–	5
Corylus avellana	–	–	–	–	–	–	–	1	1	–	–	–	–	–	1
Fraxinus excelsior	–	–	–	1	–	–	v	2	3	–	–	–	–	–	3
Total	4	12	16	3	1	9	1	17	31	1	9	3	3	16	63

). Six better-preserved shoot samples from both concentrations were anatomically examined; results indicated that, like the pointed stakes, they were made of willow-type species (3/6) and hazel (3/6). For 33 out of 48 wooden artefacts, the stage of the final annual ring formation was determined. Stage A was identified for 2 piles, stage C for 3 stakes, and stage B–C for 28 piles, stakes, and shoots (Table 1).

At Šventoji 41A-2, the piles were made of hazel (3/6), ash (2/6), and alder (1/6) wood (Table 1). Their tips had been sharpened either by chopping a flat facet (Figure 8(1)) or by removing long flakes (Figure 8(5)) and chopping around the perimeter (Figure 8(2)). For one ash pile, the final annual ring formation stage was determined as stage B (Figure 9B). Shoots from the western concentration were made of alder (3/7) and hazel (4/7), while at the eastern concentration, the species of the shoots could not be identified due to advanced decay (Table 1).

At Šventoji 41A-3, the stakes were made of alder (5/9), ash (3/9), and hazel (1/9) (Table 1). Their tips were sharpened by removing long flakes (Figure 8(3,4)). For one pile, the final annual ring formation stage was determined as stage A (Figure 9A; Table 1).

At Šventoji 41A-1, two wooden samples were radiocarbon dated: an alder pile (4111 ± 28 BP) and one willow-type (Salicaceae) stake (4150 ± 29 BP). The dates do not differ significantly (T = 0.9, T′(5%) = 3.8, ν = 1; [39]) and can be combined to 4130 ± 21 BP (2868–2584 cal BC). At Šventoji 41A-2, two piles (alder and hazel) and one hazel shoot from the western concentration yielded ¹⁴C ages of 4065 ± 35, 4120 ± 30, and 4080 ± 50 BP. These are not significantly different (T = 1.5, T′(5%) = 6, ν = 2; [39]) and can be combined to 4094 ± 21 BP (2851–2504 cal BC). At Šventoji 41A-3, two piles (ash and alder) were dated to 4155 ± 35 BP and 4125 ± 35 BP; these are statistically consistent (T = 0.4, T′(5%) = 3.8, ν = 1; [39]) and can be combined to 4150 ± 29 BP (2874–2586 cal BC). Together, the radiocarbon dates from the three fishing sites indicate an age of ca. 2850–2500 cal BC, corresponding to the beginning of the Neolithic in the eastern Baltic.

In addition to the Neolithic wood dates, we report four ¹⁴C determinations from the Subneolithic fishing stations Šventoji 42 and Šventoji 41A-4 (

Table 2. Radiocarbon dates of Subneolithic and Neolithic stationary fishing sites at Šventoji.

No.	Fishing Site	Period	Lab Code	Date BP	Cal BC/AD (95.4%)	Description, ID	Reference
1	Šventoji 1	Neolithic	Poz-197686	4255 ± 35	2923–2701	wooden (Rosaceae) handle of the stone battle axe, EM2070:510	this study
2	Šventoji 41A-1	Neolithic	FTMC-UJ17-11	4150 ± 29	2876–2627	wooden (willow family) stake, ID 158	this study
3	Šventoji 41A-1	Neolithic	FTMC-UJ17-10	4111 ± 28	2866–2574	wooden (alder) pile, ID 124	this study
4	Šventoji 41A-2	Neolithic	Poz-193160	4065 ± 35	2848–2475	wooden (alder) pile, ID 56	this study
5	Šventoji 41A-2	Neolithic	Poz-193348	4080 ± 50	2866–2475	wooden (hazel) shoot, ID 43	this study
6	Šventoji 41A-2	Neolithic	Poz-193159	4120 ± 30	2868–2577	wooden (hazel) pile, ID 35	this study
7	Šventoji 41A-3	Neolithic	Poz-193347	4095 ± 35	2866–2497	wooden (maple) handle of the stone battle axe, ID 37	this study
8	Šventoji 41A-3	Neolithic	Poz-193158	4125 ± 35	2871–2578	wooden (alder) pile, ID 30	this study
9	Šventoji 41A-3	Neolithic	Poz-192941	4155 ± 35	2881–2623	wooden (ash) pile, ID 28	this study
10	Šventoji 41A-4	Subneolithic	Poz-192943	4205 ± 30	2897–2674	wooden (alder) pile, ID 80	this study
11	Šventoji 42	Subneolithic	Tln-3415	4177 ± 60	2900–2580	wooden (alder) pile, ID 115	this study
12	Šventoji 42	Subneolithic	Tln-3421	4215 ± 60	2920–2620	wooden (alder) pile, ID 118	this study
13	Šventoji 42	Subneolithic	Tln-3449	4239 ± 60	3010–2630	wooden pile, ID 5	this study
14	Šventoji 42	Subneolithic	Vs-1956	4280 ± 40	3018–2762	wooden pile, ID 3	[23]
15	Šventoji 42	Subneolithic	Vs-1981	4090 ± 50	2872–2491	wooden pile, ID 14	[23]
16	Šventoji 58	Neolithic	Poz-77557	4000 ± 35	2619–2462	wooden (hazel) shoot of fish screen, test-pit 101, ID 18	[15]

). Three samples from Šventoji 42 do not differ significantly (4177 ± 60, 4215 ± 60, and 4239 ± 60 BP; T = 0.5, T′(5%) = 6, ν = 2; [39]) and can be combined to 4210 ± 35 BP, corresponding to 2902–2671 cal BC. These ¹⁴C ages, measured at the Tallinn laboratory, are generally consistent with two previously published determinations from the Vilnius laboratory obtained using the same method—liquid scintillation counting (4280 ± 40 and 4090 ± 50 BP) (Table 2). At Šventoji 41A-4, an alder pile was dated to 4205 ± 30 BP, corresponding to 2897–2674 cal BC.

4.2. Shaft-Hole Battle Axes

Two stone battle axes with wooden handles were examined from the Šventoji 41A-3 and Šventoji 1 fishing sites. The axe from 41A-3 has a head measuring 12.5 × 5.3 × 3.4 cm, made of dolerite (Figure 6). According to the Central European typology, it belongs to type B [40], and according to the Lithuanian classification, to the European boat-shaped type with a truncated-cone butt [41]. The handle is 54 cm long and 2 cm thick. For microscopic analysis, a wood sample measuring 5 × 5 × 5 mm was taken from the handle. In the transverse section, diffuse-porous wood with solitary, widely spaced pores were observed (Figure 10(A-1)). In the tangential section, uniseriate, biseriate, and rarely triseriate rays, not exceeding 30 cells in height, were visible (Figure 10(A-2)). In the radial section, homogeneous rays, spiral thickenings, and large ray–vessel pits were present (Figure 10(A-3)). These anatomical features suggest that the axe handle was made of Acer sp., and the relatively narrow and short rays indicate Acer campestre.

About 1.2 km south of the 41A-3 fishing site, another stone axe with a wooden handle was found in the 1960s at the Šventoji 1 site, located in the bed of the same lagoonal lake [18]. The axe measures 14.5 × 5.3 × 4.2 cm, with a shaft-hole 1.7 cm in diameter, and is made of greenish-grey amphibolite (Figure 11). By form, it corresponds to type A or the European type, but without a ridge on its back [40,41]. The handle measures 24.5 cm in length, though it was originally longer, as it appears broken. According to R.R., who did not specify the method of identification, the handle was made of juniper or yew [18]. Our analysis, however, does not confirm this. In the transverse section, numerous solitary pores were observed, more densely arranged at the ring boundary (Figure 10(B-1)). In the tangential section, mainly biseriate and uniseriate rays not exceeding 20 cells in height were observed (Figure 10(B-2)). In the radial section, homogeneous rays, simple perforation plates, and small intervessel and vessel-ray pits were recorded (Figure 10(B-3)). Based on these anatomical features, the axe handle from Šventoji 1 was identified as belonging to a species of the Rosaceae family, tribe Maleae.

Traceological analysis of the axe handles was severely limited by poor preservation and previous conservation treatments. The appearance of both handles has been substantially altered post-depositionally. The surface of the specimen from the Šventoji 1 site is cracked and sunken (Figure 12A), with only a few areas remaining well-preserved. Remnants of the preservative are visible in many places (Figure 12B). The area of the handle containing the axe is visibly shrunken and discoloured black (Figure 12C). Essentially, no better-preserved technological traces were observed on the artefact. Only the tip shows remnants of incisions resulting from whittling (Figure 12D), and similar traces, though much less well-preserved, were also observed on the central part of the handle (Figure 12E). The tip was whittled into a semi-circular shape, now evidenced only by the multi-planar nature of the surface, as other traces have been erased. Only one type of use–wear mark was observed on the handle: traces of fibre compression resulting from the use of a stone tool mounted on it (Figure 12F).

The primary result of the microscopic analysis of the axe from Šventoji 1 is that much of the surface damage visible on the artefact is post-depositional flaking (erosion) rather than modifications resulting from use. This is indicated by the scattered distribution of flaked areas, but above all by the good preservation of the raw material crystals in these regions, which show no crushing and retain varied microrelief (Figure 12G). Technological traces, aside from the intentional polish observed on the preserved surfaces (Figure 12H), were noted only around the tool’s hole, which was most likely initially marked using the nicking technique—striking with a pointed stone tool—to roughen the surface and create a slight depression for drilling (Figure 12I). After drilling, the area was ground.

The axe from Šventoji 1 displays several categories of use–wear traces. These include rounding and matting of the hole’s edge due to contact with the handle (Figure 12J). The characteristics of the polish suggest that leather may have been used to bind the axe. In addition, the axe blade shows sand-filled striations oriented perpendicular to the blade’s axis, likely of utilitarian origin (Figure 12K,L). The blade’s edge is crushed from use and probably eroded, although it is difficult to determine the full extent of this damage (Figure 12L).

At the time of analysis, the handle of the axe from 41A-3 had not yet undergone conservation and was submitted for microscopic observation in a “wet” and very brittle state. Its surface contained a large amount of dirt and bark remnants, which complicated observation (Figure 12M). Technological examination revealed that the tip of the handle had been formed by whittling and cutting (Figure 12N), and the base of the handle was shaped in a similar manner (Figure 12O).

The wooden handle of the Šventoji 41A-3 battle axe was dated to 4095 ± 35 BP, corresponding to 2866–2497 cal BC (Table 2) while the ¹⁴C age do not differ significantly from those of two piles—4155 ± 35 and 4125 ± 35 BP. These three dates are not significantly different (T = 1.5, T′(5%) = 6, ν = 2; [39]) and can be combined to give 4150 ± 29 BP, corresponding to 2874–2586 cal BC. The wooden handle of Šventoji 1 was dated to 4255 ± 35 BP, corresponding to 2923–2701 cal BC (Table 2).

5. Discussion

5.1. Construction and Function of Neolithic Fishing Stations

All three Neolithic fishing sites we investigated were located on flat sections of the former lakebed, although at different depths. Based on the topographical position of Neolithic settlements and fishing sites, the water level in the Neolithic period is estimated to have been approximately 1 m a.s.l. [17]. At Šventoji 41A-1, the remains of the shoots lay at −0.65 m a.s.l., indicating that the lake depth during the site’s use was about 1.65 m. At 41A-2 and 41A-3, the remains of fishing structures lay at −0.1 m a.s.l., suggesting shallower water, around 1.1 m deep. These depths are approximate, as water levels would have varied seasonally.

A key question arising from the new archaeological material concerns the origin of the shoot concentrations found at Šventoji 41A-1 and 41A-2. According to numerous ethnographic and archaeological records worldwide, such finds—discovered lying in the bed of the former lake together with vertical wooden piles—may represent only the remains of fish traps or fragments of fishing fences [42,43,44,45,46]. Furthermore, ten years ago a very similar fragment of a Neolithic fish fence, in much better preservation, was discovered at another site—Šventoji 58. It was composed of hazel rods measuring 1.8 m in length and approximately 1 cm in diameter, bound together by a lime bast thread [15].

However, without doubting their fishing function, it is particularly difficult to determine whether the clusters of shoots at Šventoji 41A represent remains of fishing fences or of fish traps. Unlike the Subneolithic layer, the Neolithic horizon at Šventoji 41A is located in higher and drier gyttja layers, which, after drainage works in the 1960s, dry out in summer and become waterlogged again in autumn. Such fluctuations in groundwater are highly destructive to archaeological wood, leading to extensive decay of the horizontal shoots and the complete loss of any binding cords.

This is further complicated by the high likelihood that both traps and fences could have been made in the Neolithic from the same construction elements—round shoots of fast-growing trees such as hazel, willow, and alder. For instance, during the Subneolithic period, pine laths up to 2.5 metres in length were used for the construction of both fish fences and traps at Šventoji [18]. However, these fragmentary finds do not allow us to determine the size or construction of the traps. For instance, at the Šventoji 2/4 site, within a Subneolithic layer, a possible fragment of a fish trap made of pine laths tied to a hazel rod frame was discovered (Figure 15 from [20]). Alternatively, this could represent a panel of a fishing fence or a box-shaped fish trap (Figure 183 from [18]). A five-metre-long fish trap made of pine laths has been briefly reported from the Kretuonas 1C site in northeastern Lithuania, dating to around 2000 cal BC [47]. Here, the presence of Neolithic cultures (GAC and CWC) was only episodic, while the local hunter-gatherer-fishers continued their Mesolithic way of life until the very beginning of the Bronze Age, around 1800 cal BC, or possibly even longer [48]. Therefore, the tradition of constructing fish fences and traps from pine laths continued throughout the Neolithic in more inland regions. Better-preserved and better-documented fish traps are known from Mesolithic and Neolithic sites in northeastern Russia, the Netherlands, and Denmark, typically cone-shaped with an inner “throat”, constructed from wattle, frames, and binding material [45,46,49,50,51].

At the Neolithic Šventoji sites, fish traps made of tied deciduous tree shoots could have been used, but it is likely not all shoot concentrations we observed represent trap remains. Some may have been remains of lightweight, movable fences, as in Šventoji 58 [15]. Such a fence lacks a frame, allowing it to be rolled for transportation and then attached to piles or stakes at the fishing site. Fishing fences made from bark-tied hazel rods are known from Denmark and northwestern Germany and are often dated to the 3rd millennium cal BC [44,46]. Remains of similar fences, although made from pine laths instead of round wood shoots, are known from Subneolithic horizons at all the Šventoji waterlogged sites [18], as well as at Sārnate in coastal Latvia [52], dating to ca. 3500–2750 cal BC. However, none of the pine lath sections have been found standing in a vertical position and attached to wooden piles or stakes, as they would have been in use.

At Šventoji 41A-2, the western shoot concentration had spread over about 2 m in width, which is too wide for a fish trap and more closely resembles a fence fragment of similar length. In contrast, the eastern concentration widens at one end and narrows at the other, suggesting the remains of a conical fish trap, although a similar pattern could also have been caused by the collapse of a fish fence (Figure 5). Fish traps could be used individually, tied to one or several wooden piles, or in combination with fence, while fences required additional piles for support. A group of nine piles at 41A-3 may represent part of a ≥12 m-long straight fence, whereas the six-pile group at 41A-2 may correspond to a >10 m-long curved fence (Figure 5 and Figure 8).

At Šventoji 41A-1, within and adjacent to the concentrations of shoots, numerous thin vertical stakes were recorded, which were entirely absent at the 41A-2 and 41A-3 fishing sites. Nevertheless, stakes and piles at 41A-1 form a 6 m line, likely extending ≥12 m, as two piles were found further west in a 2012 test pit. In the area of the eastern shoot concentration, stakes and piles are entirely absent, suggesting an empty section of the fence intended for a fish trap. By contrast, the western concentration contains numerous vertical stakes, indicating that the shoots may have remained from a collapsed fence. Unfortunately, the available data are insufficient to reconstruct the form of any Neolithic or Subneolithic fish weir in the southeastern Baltic.

Wood species analysis from sites 41A-1, 41A-2, and 41A-3 indicates that alder, hazel, and European ash were the most commonly used woods for making piles during the Neolithic period. Across the three sites, the proportions of these species differ, with alder predominating at 41A-1 and 41A-3 (Figure 13). However, the number of piles is too small—particularly at 41A-2 (n = 6) and 41A-3 (n = 9)—to allow for a more detailed interpretation. Previous analyses of wooden finds from Šventoji indicate that hazel, alder, and ash were the preferred species for constructing fishing structures in stagnant waters during both the Subneolithic and Neolithic periods. Slight variations in their proportions most likely reflect differences in local wood availability rather than the deliberate selection of specific species. The situation appears to have been changing during the Bronze Age, when the lakes had already become overgrown, and ash wood became the most desirable material for river-fishing structures, as evidenced at the site of Šventoji 55 [53].

Analysis of the cutting season of wooden stakes and piles shows that most construction wood was harvested during the non-vegetation period, from autumn to early spring, as indicated by fully formed last annual rings. In several cases (2/16 at 41A-1 and 1/9 at 41A-3), the last rings were only beginning to form, indicating that the wood had been cut in early spring (Table 1). Therefore, it may be assumed that most stationary fishing structures were built during the winter and subsequently repaired. However, construction could also have taken place in the warmer season, with the wood prepared in winter. Currently, there are insufficient data to compare the seasonality of wood harvesting in the Neolithic versus the Subneolithic. Interestingly, similar patterns were observed at the Šventoji 55 Bronze Age fishing site, where in 35 of 47 cases, wood was prepared during the non-vegetation period [53].

Regarding the shaping of wooden pile tips, there is no evidence of a marked Neolithic innovation or technological break. Based on visual evaluation only, it seems that the same sharpening techniques observed at Neolithic sites are also present in Subneolithic contexts, for example, at Šventoji 41A-4 (unpublished data) and 42 [23]. Piles were sharpened in several ways: to a pencil-shaped point, by removing one or several long flakes along part of the perimeter, by chopping 1–3 facets, or left unsharpened. Based solely on macroscopic assessment, it is likely that the long sharpening flakes were detached from the cut end of the trunk using wedges. However, experimental work to test this hypothesis highly needed, although it is still in the planning stage. At Šventoji, ground stone axes and adzes were already in use during the Subeolithic [19], so the introduction of shaft-hole axes in the early Neolithic (see Figure 9 and Figure 11) could not significantly affect pile and stake production.

In the Neolithic horizons of the Šventoji 41 sites, we identified two unworked and two flaked stones, all of which could have been used as net sinkers. However, their small number, the absence of net floats at the fishing sites, and the possibility that these stones may instead have been attached to fish traps or fences make it uncertain whether nets were actually used at the Neolithic fishing sites we investigated.

No fish bones were recovered from the Neolithic horizons of the fishing sites 41A-1, 41A-2, and 41A-3, and several factors may explain this absence. The bones may have completely decayed due to the periodically drying gyttja. However, it is more likely that substantial accumulations of fish bones never formed at these sites, while sparsely distributed small bones went unrecovered during the excavations, since wet-sieving of gyttja at the Šventoji sites is highly ineffective and was not employed. Fish caught on site were likely transported to nearby settlements on the shore, where they were processed and consumed. This interpretation is supported by the almost complete absence of settlement-related remains, such as pottery or evidence of amber and flint processing, suggesting that the fishing sites were used only briefly and that primary fish processing did not take place there. The only information on fish species caught in the Neolithic comes from the upper archaeological layers (A1 and A2) at Šventoji 2/4, dated to ca. 2800–2620 cal BC. These remains belonged to cyprinids, pike, and zander. Diatom data also indicate that the lagoonal lake had minimal brackish inflow at the time, meaning that mainly freshwater fish inhabited it. Fishing ceased when the Šventoji River found a new course to the sea, cutting off the lake’s inflow [20].

5.2. Dating of Neolithic Fishing Sites

The Šventoji archaeological sites have been dated using approximately 180 radiocarbon dates, making this complex one of the best-dated in Lithuania [16,17,18,19,20]. Unfortunately, due to the plateaus in the radiocarbon calibration curve, some periods cannot be dated as precisely as desired [32,54], and dendrochronological dating is impossible because the piles used were relatively thin (4–8 cm) and of unsuitable tree species (e.g., hazel, alder, ash). One such plateau (2880–2580 cal BC) coincides with the beginning of the Neolithic in the southeastern Baltic region [55] and hinders precise dating of the key technological change that marks the start of Neolithic fishing—the transition from pine laths to round deciduous shoots.

Currently, we have 16 ¹⁴C dates from the fishing sites, whose assignment to the Subneolithic or Neolithic periods, based on diagnostic artefacts such as battle axes and stationary fishing technology, is unambiguous (Table 2). Based on the obtained radiocarbon dates and the limited number of artefacts recovered, the Neolithic sites 41A-1, 41A-2, and 41A-3 appear to have been used during relatively short periods between approximately 2850 and 2500 cal BC. The youngest Subneolithic fishing sites, such as Šventoji 42 and Šventoji 41A-4, were in use between roughly 3000 and 2500 cal BC. The calibrated probability ranges of the Subneolithic and Neolithic fishing sites are broad and largely overlap; however, this does not necessarily indicate simultaneous use. On the contrary, material culture appears to have changed rapidly at the beginning of the Neolithic, likely due to the arrival of new populations. In stratified Šventoji sites (e.g., 2/4), Subneolithic porous pottery is consistently found below Neolithic GAC ceramics [17,20]. Assuming that the technological shift in fishing occurred abruptly and was irreversible, the dates of Subneolithic and Neolithic fishing sites can be modelled using the PHASE function [31], applying non-overlapping Subneolithic and Neolithic phases. This approach constrains the timing of the event to approximately 2800–2650 cal BC, with a 95.4% probability (Figure 14) and that time corresponds to the appearance of the Neolithic populations in mainland Lithuania [55,56,57,58].

5.3. Users of Neolithic Fishing Sites

Determining the identities of the Neolithic fishermen is complex for several reasons. The most limiting factor is that no human skeletal remains dating to the Neolithic have been found at Šventoji; all recovered remains are attributed to the Subneolithic [59]. Furthermore, the shores and bed of the lagoonal lake contain material from two Neolithic cultures—Globular Amphora Culture (GAC) and Corded Ware Culture (CWC) [17]. At fishing sites yielding numerous artefacts, including stylistically classifiable pottery, it is often impossible to assign specific fishing structures to either culture because the archaeological layers were unstratified or the stratigraphy was not properly recorded during earlier excavations (e.g., Šventoji 1). Meanwhile, in short-term Neolithic fishing sites such as 41A-1, 41A-2, 41A-3, 41A-4, and 42, pottery is absent, and other artefacts are very scarce.

Fortunately, at site 41A-3 we discovered a stone battle axe that can be confidently attributed to the CWC rather than the GAC (Figure 9). Stone axes with preserved wooden handles are extremely rare across Europe. Comparable examples include a polished flint axe with a wooden handle, dated to ca. 3500 cal BC, from Lolland Island [60]; a thin-butted flint axe from Sigerslev Bog, Denmark [61], and at least three stone axes with preserved wooden handles from the latter half of the 4th millennium cal BC in Great Britain [62]. However, all of these finds predate the CWC, and until now, no other examples of shaft-hole battle axes with preserved handles have been documented in Europe except at Šventoji, specifically—at Šventoji 1.

The axe handle from Šventoji 1 was made from a Maleae genus tree of the Rosaceae family, whereas the 41A-3 axe handle was made from field maple (Acer campestre). These species were not commonly used for fishing equipment during either the Subneolithic or Neolithic periods. Only isolated examples exist of maples or Rosaceae family trees used for fishing constructions and tools; most were instead made from alder, hazel, ash, or willow—species that dominated the lakeshore environment. Because Rosaceae and maple trees were rare in the immediate vicinity of the lake and grew farther inland, they were therefore selectively chosen for specific tools [53]. Species of the Maleae genus, such as rowan, have dense, durable wood suitable for tool handles [63], while forest apple (Malus sylvestris) provides a hard, uniform, and aesthetically appealing material ideal for carving [64]. Thus, the selection of these woods for axe handles appears deliberate and functionally motivated.

The choice of field maple for the 41A-3 axe handle raises further questions. Maple wood used in Subneolithic and Neolithic constructions at Šventoji typically exhibit broad and tall rays, characteristic of Norway maple (Acer platanoides). For example, a decorated Subneolithic paddle made of Norway maple was found at Šventoji 5 in 2021 [65]. In the tangential section of this paddle’s wood, 30–50 cell-high, five- to six-seriate rays were observed—an anatomical features characteristic of Norway and sycamore maples. Presently, Norway maple is the only naturally occurring maple species in Lithuania. According to the European Forest Genetic Resources Programme [66], the northern limit of the modern distribution of field maple lies approximately 200 km south of the Šventoji sites, in northern Poland.

Several scenarios could explain the use of field maple in the 41A-3 axe. If its Neolithic distribution was similar to that of today, the axe and its handle likely originated from the south. Alternatively, the northern range of field maple may have extended farther north during the Neolithic period, allowing it to grow near the Šventoji lagoon. However, since pollen analysis cannot distinguish between Acer species, this hypothesis remains untestable at present.

The Šventoji 1 site contains a substantial assemblage of CWC pottery, comprising at least 33 vessels. These potsherds were found within the same horizon (A) as GAC pottery, which represents a minimum of 55 vessels [55]. Nevertheless, there is no doubt that the boat-shaped battle axe from Šventoji 1 belongs to the CWC context, as neither the Rzucewo Culture nor the GAC groups produced such forms [58,67]. In addition to the two axes with preserved handles discussed here, two more battle axes have been recovered at Šventoji: one from the lake shore [55] and another from the lakebed near an ornate amber disc [18]. The context of the latter find suggests that at least some of these axes were not accidently lost or deposited in graves but rather were intentionally deposited in ritual contexts within the lake.

Both battle axes recovered from the fishing sites were small (12.5–14.5 cm in length) and had thin wooden handles (1.7–2 cm Ø), making them ineffective for felling trees but potentially suitable for other woodworking tasks, such as serving as wedges for sharpening piles. These artefacts raise questions regarding their function: were they weapons and symbols of CWC warrior status [68,69] that were lost or ritually deposited at fishing sites, or were they important tools used in the manufacture and maintenance of fishing equipment?

Unfortunately, the state of preservation of the artefacts prevented a comprehensive examination, and at this stage of the research, no definite conclusions can be drawn regarding their specific function. Firstly, the diameter of the holes in many other known CWC axes is similar, which prevents the handle diameter from being considered as reliable indicator of a particular use. Secondly, the blades of both tools are rounded and wide-angled. Experimental studies conducted by the authors (results to be published in a separate article) suggests that these characteristics do not favour their use for splitting wood, particularly thinner branches, although such use cannot be entirely excluded. The crushing and linear marks visible on the blade of the Šventoji 1 axe are almost certainly unrelated to woodworking; they are too abrasive and most likely resulted from contact with a harder material, possibly stone. It cannot, however, be excluded that these traces were caused by accidental impact with stone during pile sharpening. Traces of use on the analysed handles, together with blade damage on one of the axes, indicate the long-term use of these artefacts.

Research at Šventoji 41A demonstrates that CWC people engaged in stationary fishing at the Šventoji lagoon. Neolithic fishermen exploited both previously used Subeolithic fishing sites (e.g., Šventoji 1 and 6) and new locations (e.g., Šventoji 41A). The 41A fishing sites were situated approximately 200 m from the eastern lake shore (Figure 1C), which was low and wet. No CWC settlements are known along this shore, suggesting that the fishermen’s dwellings were located farther from the fishing sites—either on the sandy spit on the western lake shore (approximately 1.2 km away), on Šventoji 1 island (1.2 km), or near the Šventoji River, for example, at Šventoji 40 (1.2 km) (Figure 1A).

5.4. Concerning De-Neolithisation in the Southeastern Baltic

How does the discovery and investigation of Neolithic fishing sites at Šventoji—belonging to the CWC, and possibly also to the GAC—change or support our current understanding of the Neolithisation process in the southeastern Baltic region? Firstly, our results provide new and robust evidence that representatives of the CWC engaged in fishing along the Lithuanian coast. This is not entirely unprecedented, as previous studies had already indicated a significant freshwater contribution to the CWC diet, based on carbon and nitrogen stable isotope analysis of bone collagen and on food residues found in ceramic vessels [15,16,70]. Secondly, our research indicates that CWC fishermen engaged specifically in stationary fishing at Šventoji, a finding that is both new and significant for understanding the settlement systems and subsistence strategies of this culture. Fish weirs provide a larger and more stable catch compared to active fishing methods, but they require constant maintenance and repair, as well as the regular removal of fish from the traps [43]. Consequently, engagement in stationary fishing would have undoubtedly reduced the mobility of fishermen. Elsewhere in Europe, fish weirs attributable to the CWC are rare. At Syltholm Fjord in Denmark, a notable decline in stationary fishing has been observed, dating to ca. 3000–2550 cal BC [46], which corresponds to the earlier part of the CWC period [71]. Thirdly, the study of Neolithic fishing sites at Šventoji indicates that CWC people constructed fish weirs and traps differently from Subneolithic fishermen (Figure 15). They used different materials and most likely employed distinct production techniques. This suggests that, upon arriving in the eastern Baltic from Central Europe—as supported by aDNA data [3,72,73]—the CWC people retained their cultural identity and specific practices not only in pottery production and stone processing, but also in stationary fishing. CWC fishermen did not adopt freshwater fishing techniques from local Subneolithic populations, but instead brought their own knowledge and methods, or possibly learned them from GAC groups, who likely entered the region slightly earlier and also engaged in fishing [16,18,58].

But can this evidence of freshwater resource exploitation by Neolithic communities be interpreted as a sign of de-Neolithisation? In the case of the GAC, the answer is most likely yes, since in their homeland—Central Europe—they lived as devoted farmers [74,75] and had to adapt their lifestyle and dietary preferences upon reaching the Baltic coast. However, for the CWC, the answer is more complex, as evidence—mostly from coastal areas—indicates that hunting and fishing were incorporated into their subsistence strategies to a significant extent [76,77]. Furthermore, elevated δ¹⁵N values in CWC individuals may suggest that freshwater fish contributed substantially to their diet, even in Central Europe [78,79]. The fish weirs associated with the CWC at Šventoji could be considered an indicator of de-Neolithisation in economic terms only under two conditions: either if the region had already undergone Neolithisation prior to the arrival of CWC groups, or if the earliest CWC communities settling there were strictly agrarian. At present, however, no evidence supports either of these conditions. Accordingly, this phenomenon is better interpreted not as a reversion to pre-Neolithic subsistence, but as an expression of the economic flexibility and diversity of the CWC, in which farming and animal husbandry were integrated with the exploitation of aquatic and forest resources.

6. Conclusions

Recent investigations of Neolithic fishing sites at the Šventoji archaeological complex in northwestern Lithuania not only confirm that Neolithic communities practised stationary fishing in a lagoonal lake, but also support a previously proposed hypothesis that, around 2800–2650 BCE, an important technological shift occurred in stationary fishing. The main construction elements of wooden weirs and fish traps became straight and thin shoots of fast-growing deciduous trees (willow, hazel, alder), replacing the pine laths used in the Subneolithic. The sites, characterised by this new type of fishing fences and traps, were likely used by Corded Ware Culture (CWC) communities, as evidenced by the discovery of a battle axe at one of them. The role of battle axes in building and repairing wooden fish weirs remains unclear and warrants a dedicated experimental study. Finally, the Neolithic fishing sites at Šventoji further illustrate the economic flexibility and diversity of the CWC in the eastern Baltic, where animal husbandry was integrated with the exploitation of aquatic and forest resources.

Additionally, the two shaft-hole stone battle axes with preserved wooden handles found at the Šventoji Neolithic fishing sites (1 and 41A-3) remain the only such finds in Europe to date. Unfortunately, the state of preservation of both artefacts prevented a more comprehensive use–wear analysis. Only a few technological details could be observed regarding the production of the wooden handles and the stone axes themselves. Traces of use on the analysed handles, along with damage visible on the blade of one of the axes, indicate long-term use of these artefacts. Nevertheless, it was not possible to confirm a direct functional relationship between these axes and the prehistoric fishing structures described in this article.