The coastal spread in the western Mediterranean was much faster than in central Europe, which may be explained by long-distance maritime travel and exploration combined with demic expansion in coastal areas and assimilation of local foraging communities (Rigaud, Manen, and García-Martínez de Lagrán 2018). Through the Mediterranean, different chronological and geographical traditions appear.
The Impressa ware develop around 6000 BC in the Ionian Sea, and constitutes the first step in the expansion to the west. It replaces the monochrome pottery around the southern Adriatic, appearing first in west Greece and south-eastern Italy, and spreading fast to the north into Dalmatia, and to the west into southern Italy and Sicily (Suppl. Fig. 6). The primary distribution of the “leap frog” type of impressed ware led by small pioneering units with a well-established agropastoral economy is seen rapidly expanding in western Mediteranean sites ca. 6000–5600 BC, such as Sicily, Tuscany, Liguria, Provence, Languedoc, Valencia (Guilaine 2017).
The common material culture element is thus pottery impressed with diverse instruments, such as the Cardium edule shell in the western Mediterranean, from the Adriatic coast to Portugal. A secondary phase of generalisation and regional settlement starting ca. 5600/5500 BC sees the development of specific Early Neolithic cultures, such as Stentinello in Sicily and Calabria; Tyrrhenian Cardial in Latium-Tuscany-Sardinia-Corsica; regional groups of the Franco-Cantabrian Cardial (Provence, Catalonia, Valencia, Andalusia). Cardial groups spread also into Morocco, either from Sicily or Iberia (Guilaine 2017).
In Iberia and southern France, agriculture spreads rapidly as flora diversity increases with the arrival of farming/herding communities, transforming the – until then – intact temperate forest areas. For hundreds of years a clear increase in demography, technological developments and cultural practices unfold, and during the Middle Neolithic the proportion of dominant species changes to those resistant to constant cuts (deciduous Quercus, evergreen Quercus, Olea, Pinus halepensis), and open areas multiply. Cycles of land exploitation, associating wood cutting, farming and herding, followed by woodland regeneration are common (Guilaine 2017).
While economic practices changed, some cultural traits like funerary practices did not accompany the ‘Neolithic package’ acquired in the Balkans by local fisher-hunter-gatherers. It seems that farmers and their domestic animals spread fast, in ca. 10 human generations, from sub-Mediterranean Macedonia to the northern limits of the temperate Balkan Peninsula and the adjacent Carpathian Plain, which may have put serious difficulties for the spread of cattle until selective pressure could provide genetically-driven adaptations to harsh environments (Ethier et al. 2017).
The early 6th millennium pioneer settlers in the interior of the Balkans were probably the first to face these challenges. Among their response to unfamiliar ecological conditions, the easiest adaptation was to adjust the species mix in favour of crop and livestock taxa that reproduced best in the new environment. However, farmers from the southern Balkans (modern-day Bulgaria and northern Macedonia) chose a strategy of diversification, exploiting a very broad spectrum of crops, probably to reduce climate-related losses. With their expansion to the northern Balkans and the Carpathian Basin, farmers abandoned many leguminous crops, and reduced the spectrum of cultivated plants (Ivanova et al. 2018).
On the other hand, the faunal assemblages were clearly structured by climate: sheep, goats, and pigs were reduced, while cattle and wild spaces increased in frequency during the northward expansion of farming. This, in combination with dietary evidence – such as organic residues in ceramics and stable isotope values in human bones – suggests that animal husbandry and especially dairying became of key importance for the initial establishment of farming beyond the Mediterranean areas (Ivanova et al. 2018).
The Linearbandkeramik (LBK) culture brought (starting ca. 5700–5600 BC) the first farming settlements to central European uplands as well as the North European Plains along the Oder and Vistula rivers. It expanded by colonising habitats favourable to agriculture, through a progressive migration of farming peoples from the Danube Valley to the north and west. Its economy was based almost entirely on domesticated plants and animals, and settlements were concentrated on fertile loess soils along streams. Analysis of isotope in skeletons in the Rhine Valley suggest that peoples of local origin may have been involved in the establishment of these early farming communities (Midgley 2004), which suggests short-range migrations.
Early Neolithic warfare in central Europe shows violent organised conflict between independently acting (probably territorial) groups connected by kinship ties. Massacres seem to cluster near the end of the LBK period, in the decades before 5000 BC, which suggests profound changes that affected interlinked social and natural landscapes: climate-induced drops in agricultural production with mounting claims to inherited agricultural land and increasing hierarchical differentiation are likely factors for the rise of social tensions and lethal conflicts between local groups. Evidence of perimortem collective lethal violence, as well as possible torture, mutilation, execution, dismemberment and cannibalism, including children, shows violence was a major societal issue for later LBK populations (Meyer et al. 2018).
Neolithic farmers from Europe derive from a single Balkan population closely related to north-western Anatolians which split in two routes: one related to Danubian populations, represented by the Linearbandkeramik complex from central Europe; and another associated to the Impressa complex of Croatia and Epicardial Early Neolithic from Iberia. However, this north-west Anatolian Neolithic (NWAN) ancestry is distinct from the central Anatolian source found in Aegean Neolithic samples, being shifted away from CHG and towards WHG compared to them (Mathieson et al. 2018).
Neolithic populations from the Balkans during the 6th millennium BC cluster closely with north-western Anatolian Neolithic individuals, deriving most of their ancestry from them (ca. 98%) and the rest from WHG, consistent with archaeological evidence. An exception is found in individuals of the mid–6th millennium BC from Malak Preslavets, in the west Pontic area south of the Danube, where higher contributions of WHG (ca. 15%) and EHG (4%) are found, possibly representing populations close to the highest density of hunter-gatherers (Mathieson et al. 2018).
The Iron Gates zone represents a region of interaction between groups in both ancestry and subsistence strategy, based on strontium and nitrogen isotope data: two individuals from Lepenski Vir (ca. 6200–5600 BC), of entirely NWAN ancestry, were migrants from outside of the region and ate primarily terrestrial diet; another (ca. 6070 BC) had a mixture of NWAN and hunter-gatherer-related ancestry and consumed aquatic foods; and a fourth, earlier individual from the same site (ca. 7850 BC), had entirely hunter-gatherer-related ancestry. Another individual from Padina (ca. 5950 BC) also shows a mixture of NWAN and hunter-gatherer-ancestry, confirming the approximate date and region of interaction of both groups (Mathieson et al. 2018).
Hunter-gatherer ancestry of expanding farmers is more similar to eastern WHG individuals—like Villabruna (ca. 12000 BC) and Körös (ca. 5700 BC)—in the east, and more similar to western WHG individuals—like La Braña 1 (ca. 5900 BC) and Loschbour (ca. 6100 BC)—farther west, which shows that their admixture derives from populations with which they lived in close proximity. In particular, LBK individuals show a greater affinity to Loschbour hunter-gatherers, whereas Iberian Early Neolithic populations have La Braña-related ancestry (Lipson et al. 2017).
The increase in hunter-gatherer ancestry after the Early Neolithic period is lower in Hungary than in LBK and Mediterranean farmers from Iberia, and closest to the more eastern WHG individuals (from Villabruna and Körös), with limited intra-population heterogeneity, which points to the relative isolation of this group, close to the original source of expanding LBK farmers. In Iberia, the average admixture date is estimated ca. 5650 BC, but probably closer to ca. 5900 BC when considering only the oldest individuals assessed, which suggests—given the start of farming in Iberia ca. 5500 BC—the presence of a small proportion of hunter-gatherer ancestry in earlier Cardial Neolithic populations acquired along their migration route (Lipson et al. 2017). El Mirón-like ancestry is found in Early Neolithic individuals in higher proportions than those outside Iberia, and especially to the south, suggesting a north–south cline of admixture with hunter-gatherers who carried mixed Upper Palaeolithic ancestry (Villalba-Mouco et al. 2019).
Y-chromosome haplogroup G2a2-CTS4367 becomes also increasingly mixed with I-M170 lineages with time, especially in the Iberian Peninsula. G2a2a-PF3147 lineages (formed ca. 14900, TMRCA ca. 9700 BC), are found in Anatolian Neolithic samples from Tepecik and Barcın (and also later in Maikop from Novosvobodnaya), also early in Croatia Impressa (ca. 5560 BC), and later in Iberia, but especially in LBK samples. Haplogroup E1b1b1a1b1-L618 (formed ca. 10000 BC, TMRCA ca. 6100 BC) is also found in Cardial from Croatia (ca. 5900 BC), in Iberian Epicardial (ca. 5000 BC), and in Late Neolithic Lengyel (ca. 4740 BC), which supports its presence—probably originally from a Levantine source—in expanding Anatolian farmers.
Haplogroup G2a2b-L30 (formed ca. 14900 BC, TMRCA ca. 12500 BC), present in Anatolian Neolithic groups from Boncuklu and Barcın, as well as Greece Neolithic, seems to be more restricted to south-eastern European groups (although it is also present in the Iberian Chalcolithic), which suggests different bottlenecks during the expansion of Early European farmers from an original population with more varied male lineages.
The two major Neolithic migration waves of Neolithic settlers, from Danubian and Mediterranean routes, did probably encounter in the Paris Basin LBK and Cardium groups during the Early/Middle Neolithic transition. Cultural exchanges are observed in the region at the end of the Early Neolithic between north-eastern LBK-derived cultures Rubané Récenet du Bassin Parisien (RRBP) and Villeneuve-Saint-Germain (VSG), and Cardial farmer groups from southern France. This interaction is supported by the admixture found in ancient mtDNA from the Gurgy necropolis (Rivollat et al. 2015). Resurgence of hunter-gatherer ancestry is also seen in Middle Neolithic samples from France (Brunel 2018).
Population continuity is observed in ancient samples through the British and Irish Mesolithic, characterised by WHG ancestry sharing high genetic drift with contemporaries from France and Luxembourg. This continuity is broken with the arrival of agriculture, introduced after a millennium–long lag between the establishing of farming in the mainland by incoming continental farmers, evidenced by a massive shift in population structure, with little evidence for local admixture. There seems to be two geographically distinct entries of Neolithic farmers, the main one through the Mediterranean route of dispersal entering from north-western mainland Europe, given the more immediate affinities with Iberian Neolithic individuals (Brace et al. 2018; Cassidy 2018; Kador et al. 2018). This affinity of Middle Neolithic samples from Britain to Iberia may be explained as from a source close to La Braña and El Mirón, but also to other intermediate unsampled regions in western Europe with higher Goyet-like contribution (Villalba-Mouco et al. 2019).
While farmers and hunter-gatherers lived in settlements in close proximity during the Neolithic in the Balkans, in western, central and northern Europe, there are signs of long periods with minimal admixture (Mathieson et al. 2017). During the Middle Neolithic, a resurge of male-biased hunter-gatherer ancestry is seen in central Europe and Iberia. Persistent frontiers between hunter-gatherers and farmers are found in central and northern Europe, coincident with the loess belt of the northern European plain, to the north of which early farming techniques were probably not suitable. It is likely that new climates and environments led to the eventual breakdown of demic diffusion, and the spread of Neolithic traits by cultural diffusion (González-Fortes et al. 2017).
That resurgence of hunter-gatherer ancestry, with a ca. 4:1 WHG:EHG contribution, is found in the Balkan Neolithic in the territory of present-day Bulgaria, close to the Danube river. This suggests a heterogeneous landscape of farmer populations with different proportions of hunter-gatherer ancestry during the early Neolithic, probably due to pockets of hunter-gatherers surviving close to the coast and major rivers (Mathieson et al. 2018). There is no sign of increasing WHG ancestry in Britain as the Neolithic progresses, which discards a resurgence of hunter-gatherers (Brace et al. 2018), although there is evidence for local Mesolithic survival and introgression in southwestern Ireland, long after the commencement of the Neolithic, also implied in haplotypic-analysis (Cassidy 2018).
In Iberia, the resurgence of hunter-gatherer-related ancestry after 4000 BC occurs in higher proportions in groups from the north and centre, and is closely related to later north-western (Canes1-like) hunter-gatherers than to the El Mirón-like hunter-gatherers from the south-east (Olalde et al. 2019).