Projectile point diversity over time and space. Fishtail points are the earliest widespread projectile point type of South America (13 − 11 k cal BP) and the only one with distribution in all latitudes and in most environments, with lowest density in the tropical lowlands and highest in the Southern Cone (Figs. 1 and 2). The higher densities are observed in the areas dominated by open environments of steppes or savannas at the end of the Pleistocene, especially Pampas and Patagonia [42], where greater diversity and abundance of megafauna were also recorded[43]. Assuming that FPP were designed specially and used for hunting extinct megafauna [22] we expect, from a chronological and geographic perspective, the FPP to be replaced by other types of projectile points after extinctions all over their dispersion area (ca. 11 k cal BP).
The expectation of FPP replacement after megafauna extinctions is fulfilled for the entire Southern Cone, in general, and for the three regions (southern Andes, Pampas and Patagonia), in particular (Fig. 2). In the three regions, FPPs appear around 13 k years cal BP, just before the growth in number and density of megafaunal species stopped abruptly and began a marked declining between 12.9 k and 12.8 k years cal BP [22]. From 12 k years cal BP onward, other types of projectile points appeared in the Andes (Tuina) and Pampas (Tigre and Pay Paso) which completely replace FPP around 11 k years cal BP, at the time than megafauna almost disappeared. In Patagonia, although FPP disappear earlier than in the other regions (shortly before 11.5 k years cal BP), there is a period of around 500 years with no records of standardized projectile points. It is interesting that, in this region, after megafauna went extinct, the guanaco also retreated [44, 45] and the deceleration and decline in the growth of the human population seem to have been more drastic and longer than in other regions [23]. So, the prolonged invisibility of projectile points in the area could be also an effect of the low human population density and/or less availability of mammals after extinctions. Only around 11 k years cal BP middle-sized triangular-shaped projectile points appear in Patagonia which will persist in much of the region for several millennia (Fig. 2) [46].
Although FPP are the earliest and most widespread projectile point in southern South America, they are not the only ones of latest Pleistocene age. There are several other types contemporary with megafauna, but with a more restricted spatial and temporal dispersion than FPP (Figs. 1 and 2). Firstly, Tigre points, with a restricted dispersion in the Pampas of Uruguay, show a partial chronological overlap with the FPP and disappear together with them at ca. 11 k years cal BP. In a recent paper it has been proposed that non-stemmed triangular points are probably contemporaneous with FPP in the Pampas [47], but we have not analyzed them here because they have not yet been securely dated. The only available dates (ca. 8 k years Cal BP) come from Arroyo Seco 2 (Pampas) but these points were associated with human remains mainly embedded in the skeletons48 and so, they were probably not used for hunting or at least they weren't used exclusively for hunting. Conversely, the increase of the Tigre points temporally coincides with the decline of FPP (12 k − 11.1 k years cal BP) [49] (Fig. 2). So, it cannot be ruled out that they represent a local variant of points also designed for megafauna hunting, and that they fell into disuse and disappeared, as did FPP, after late Pleistocene extinctions.
A somewhat different case is that of the Tuina projectile points that spread over the southern Andes, which could have been partially overlapped in time with FPP (and with megafauna); although their early dates must be taken with caution, the zoo-archaeological contexts are the expected for these types of points. Tuina points, unlike Tigre, progressively increased in frequency during the early Holocene, after most of megafaunal species had already gone extinct and after FPP disappeared (Figs. 1 and 2). This chronological trend fits well with the archaeological data which show that Tuina points are usually associated with living species, especially Lama vicugna [50]. So, Tuina projectile points should not have been necessarily related to megafaunal hunting, but probably were used for hunting still extant camelids, both before and after Pleistocene extinctions.
Another latest Pleistocene projectile type, in addition to the Tigre and Tuina, but located outside the study area, is the El Jobo lanceolate points, with a more northern and Andean distribution than the FPP. Although ostensibly well-dated, El Jobo points are few and the oldest date from the Taima Taima site (ca. 16 k years cal BP [51]) is controversial [23]; they were probably penecontemporaneous with FPP. Beyond this chronological issue, the association between El Jobo points and megafauna [mastodon and megatherium] [52, 53] is strong and, hence, they could have been a local typological variant of spear heads used for killing this kind of prey. Nevertheless, since El Jobo points have only slight geographical association with Pampean megafaunal species and overlap much more with northern species, especially proboscideans, than FPP, it is not clear whether El Jobo and FPP were synchronous geographic variants of projectile points with similar function, or if the former were replaced by the latter. As experimentation by Frison [54, 55] showed, Clovis points (similar to FPP) effectively penetrated elephant hide, and Agate Basin points (more similar to El Jobo [56]) seemed better for punching through bison or cow hide. If so, effectiveness could explain replacement of El Jobo by FPP.
Morpho-functional properties of early projectile points. If FPP was the key technological factor for humans to successfully hunt megafauna beginning at 13 k years cal BP, we expect, from a morpho-functional perspective, that these points should show the highest levels of functional performance. This is because larger animals are usually more difficult to hunt and killing them, and requires more efficient and lethal weapons than those used for smaller game [57, 58]. The main function of the projectile point is to open a wound in the skin and tissue of the animal, to penetrate deep into the prey causing massive bleeding and reach some vital organ that causes rapid death, or cause serious injuries that allow pursuing and finish off the wounded animal [33, 58].
Our results show that the oldest types of projectile points of South America, in general, and FPP, in particular, are bigger, more harmful and lethal, and technologically more sophisticated than most of the early Holocene designs. As displayed by Fig. 5, the FPP from the Pampas and Andes, and the Tigre points from Uruguay, all of them temporally and spatially associated with megafauna, are those with the highest values of TCSP and, hence, highest potential capacity for tissue damage. The magnitude of the damage is mainly related with the maximum width of the projectile point, since increasing it causes a greater shock and the prey bleeds out faster [57]. Although the TCSP index is inversely correlated with the penetration capacity and reducing the size of the projectile point would increase penetration, this is limited by the need to cause injuries in the prey for rapidly bleed out [59]. Therefore, it is essential to achieve a balance between achieving a TCSP small enough for adequate penetration, and large enough to produce a wound that bleeds the prey [37, 59].
The Andes and Pampas FPP also offer a more efficient balance between width and thickness (represented by least squares line in Fig. 3). As the point thickness decreases reduces its robustness, and increases its fragility, this make it less resistant to the effects of mechanical forces acting at the moment of impact on the prey, such as buckling. Thus, as Buchanan and Hamilton [33] have remarked for Paleoindian projectile points, it seems that FPP were designed to be long, wide and thin, striking a balance between maximizing penetration power while minimizing the risk of the spearhead breaking [33]. Similar implications emerge from wound surface area analysis (Table 2) which clearly show that the FPP technology is highly effective design with a high damage capacity. In general, typologies prior to 12,000 years Cal BP would generate wounds 12–24% larger than Early Holocene ones.
By comparing the latest Pleistocene projectile technologies on a continental scale, which exceeds our study area, three additional trends can be observed. Firstly, that El Jobo points, the potentially oldest South American projectile technology, and one of the few types strongly associated with megafauna, have lesser damage capacity than southern Tigre and FPP types (Fig. 5). If El Jobo was indeed the oldest South American projectile point type, they were probably used for hunting megafauna until being replaced by FPP, probably a more effective projectile point for larger mammals hunting [54, 55]. Secondly, the North American projectile points more clearly associated with megafauna (e.g., Clovis) would have had also larger potential for tissue damage compared to later North American projectile types (e.g. Folsom) (Fig. 5 [57]). This pattern is similar to that observed in southern South America, but the values of damage capacity of the North American points are less than those estimated for Tigre and FPP. Finally, variations in size and TCSP of projectile points could be related to different weapon systems (Table IV [36]) and, as shown by Fig. 5, most of the latest Pleistocene and early Holocene types from both North and South America were probably used with atlatl darts [36], as also proposed by experimental studies [60]. Only the largest types from the Pampas (Tigre but specially FPP) seem compatible with throwing or thrusting spears [61, 62]. The latter weapon system is usually associated with hunting of larger prey [63], and in the region where megafauna is higher in frequency and diversity.
Although it has been proposed that paleoindian projectile points could have been actually knives [12, 64, 65], experimental studies of fracture shows that Fishtal points were mainly used as projectiles [60]. Moreover, the fact that the primary function of a type of point is a projectile does not imply they do not have been used as knifes [66].
In addition to the fact that latest Pleistocene projectile points from southern South American offered the highest levels of performance, their techno-functional properties also required a high degree of expertise and manufacturing cost (Fig. 3). FPP flintknappers produced points with great width and minimal thickness, using techniques including bifacial thinning, overshot flaking and usually fluting, which must have demanded a larger cost of work and higher skills compared to that needed for producing any other projectile design [7]. Points with wide and thin blades would cause more tissue damage to prey, which could have compensated for the greater work and skill invested in their manufacture. It seems reasonable to us that such a specialized weapon would have replaced previous and less efficient designs (e.g. El Jobo), but would have no longer been produced once it became an unnecessarily expensive item after the prey for which it was designed (megafauna) disappeared. In fact, most of the early Holocene projectile point types, in addition to having less damage capability, demanded lower manufacturing cost and flintknapper´s skills than those necessary for producing FPP. As displayed in Fig. 3, the work investment in Ayampitín, Pay Paso, Patagonian triangular-shaped and Tuina projectile points was limited to bifacial reduction/thinning, and smaller blade width production [67–70].
Our expectation for higher hunting performance of FPP compared to other designs are generally fulfilled, except in two specific cases. The first one refers to the Tigre points which present, together with the Pampa FPP, the highest damage capacity values. This, along with the fact that Tigre points coexisted with megafauna and disappeared right after they went extinct, suggests these points could have been also used to hunt megafaunal species. However, Tigre points are thicker and have a lesser penetration capacity than FPP (Fig. 3), as they present a less efficient relationship between width and thickness. They probably required less manufacturing costs and flintknappers skills than FPP due to the lesser work of thinning and the absence of fluting. Although the Tigre point might be only a local stylistic variant of latest Pleistocene points, its increase in frequency temporally coincides with the decline of FPP (12 k − 11.1 k years cal BP) [49, 69] (Fig. 2) and with the decreased of megafaunal abundance and diversity after 12 k years cal BP. So, an alternative explanation may be that the change in prey availability after 12 k cal BP made high performance of FPP unnecessary and Tigre points a more viable (and less expensive) technological alternative for hunting both megafauna and smaller prey.
The decreasing size of potential hunted prey in southern South America from latest Pleistocene to early Holocene is reflected in the decreased average body mass of mammals from archaeological sites (Fig. 4). If early projectiles points types varied in concert with this change, not only is the capability for tissue damage of points expected to decrease after megafauna went extinct, but also that the magnitude of this decrease ought to be consistent with the magnitude of the body mass decrease within each region. As shown by Fig. 6 our results support this hypothesis since the reduction in the TCSP median values correlates with the decrease in the median body size of the regional fauna. The differences in TCSP values are significant or marginally significant for the Andes and Pampas, but not for Patagonia, such as would be expected from the observed pattern in our results. Patagonian FPP show a relatively low damage capacity, and also a low efficient adjustment between width and thickness, even when compared against the Patagonian early Holocene points (Patagonian triangular-shaped). Even so, this pattern is not necessarily contrary to expectations because the change in the body mass of prey after extinctions in Patagonia seems to have been not so marked, entailing only the small differences in mean size of prey between periods. This is probably because latest Pleistocene Patagonian megafauna were relatively smaller compared with the Pampas and Andes, and also because early Holocene potential prey (Lama guanicoe and Hippocamelus bisulcus) were relatively bigger than those from the Andes or Pampas.
In conclusion, since the early 20th century, Fishtail projectile points have been inextricably linked with the debate over the earlier peopling of South America; mainly because of their early chronology, their morphological affinity with Clovis points and their temporal overlap with latest Pleistocene megafauna. Unlike in North America, the scant evidence for human exploitation of megafauna in South America has virtually nullified the debate on the relationship between humans and latest Pleistocene extinctions. Although a recent study showed a strong correlation between extinctions with rise of Fishtail points and inferred human population [22], no evidence of functional correlation between Fishtail points and megafauna hunting is yet available. Here we have shown, first, that changes in morpho-functional properties of early South American projectile points varied in space and time in relationship to the body size of the hunted megafaunal prey. Second, that Fishtail points were not only temporally and spatially associated with extinct megafauna, but also that they were the most efficient technology to prey on them due to their greater capacity for tissue damage, damage / penetration ratio, and lethality. Third, that greater efficiency of Fishtail points also made them technologically more expensive (due to bifacial thinning, overshot flaking and usually fluting). Fourth, that after the extinction of megafauna, Fishtail points were replaced in the different regions of southern South America by other types of points that had less damage capability, damage / penetration ratio, and lethality but were technologically less expensive, even though probably efficient enough for killing the smaller available prey. Finally, although direct archaeological evidence of human exploitation of megafauna remains elusive in South America [19], the close association in time, space, and morpho-functional characteristics between Fishtail points and the changes in species composition after Pleistocene-Holocene transition, reinforces the hypothesis that humans had a direct and significant effect in the latest Pleistocene megafaunal extinctions. Future work is needed to clarify how resharpening affected functional properties of FPP, and to evaluate potential association between hunt strategies and weapon systems and the hunting of different kind of prey (in size and ethology).