Brooks and colleagues (2005) describe evidence for distance weaponry from late MSA contexts in eastern and southern Africa. They discuss the size of points from Aduma, Ethiopia, and ≠Gi, Botswana, which are small enough that the most credible explanation is hafting on projectiles such as small thrown spears, atlatl darts, or (probably less likely) arrows.
The article begins with some consideration of why African MSA should show a pattern of regional variability in point morphologies. Explaining variability on a regional scale first requires an explanation of uniformity on a local scale:
Projectile armatures must be able to replace broken armatures, so the haft into which they are placed also imposes limitations on projectile size and form. Having a system of exchange within social networks further encourages similarity of point form so that the product remains interchangeable within a cultural group. Wiessner (1983) argued from ethnographic data that even where forms were extremely limited by raw materials regional styles emerged (e.g., fence wire in Wiessner's ethnographic example). Men hunt well for only a few years but continue to make arrow points for most of their lives; consequently, arrows are frequently given to hunters in exchange for a claim on the evntual kill. Hunting success is influenced by the hunter's familiarity with an armature; this kind of social trade network would therefore be expected to encourage homogeneity in the size and shape of projectile points. Thus social organization, especially the development of exchange networks, constrains point styles and creates sharp discontinuities at social boundaries, whether these boundaries are linguistic, ethnic or simply a result of an empty buffer zone between group ranges (Brooks et al. 2005:234-235).
This is important: if local forms were not standardized, there would be no way to recognize regional variations. Instead, we would see great variability within local samples, with relatively little additional variability between regions -- and much of that regional variation might be attributable to raw material differences.
It is worth noting that this passage makes a functional argument for local standardization, not a purely informational one. One reason to emphasize the functional argument is that regional variability is most notable for points, and less so for other elements of regional assemblages.
They compare the points from two sites, Aduma (here, considered older than 70,000 BP) and ≠Gi (with MSA levels at around 77,000 BP), to those from Tabun, Israel (through layers B, C, and D overlapping the MSA sites). The pattern of change in point dimensions at Tabun suggests that the hafting requirements were in fact a constraint on point dimensions -- in particular, the points are very conservative in the width of their bases, while the length of the point and angle of the point change over time. Against this pattern, the Aduma points decrease in thickness and width over time, but maintain the same point geometry. Brooks and colleagues suggest that the point geometry is the more functionally constrained element in the Aduma sequence, while the hafting requirements constrained the Tabun points.
Also, the weight of points decreased over time through the Tabun, Aduma, and ≠Gi samples, with the decrease most marked in the African sites. Brooks et al. relate this decline in point size and weight to the balance requirements of projectiles -- smaller projectiles demand smaller points. Also, the increased use of obsidian at Aduma tends to decrease weight further, as it is "ligher than the quartzite and silcrete of the ≠Gi points" (p. 248).
They suggest the possibility of contact between the populations that generated the Aduma and Tabun assemblages:
Marginal retouch is another area in which Aduma MSA points are distinguished from Tabun points. Each piece was divided into six retouch areas, three per side, and the presence, position and nature of the retouch, if any, were noted fore each area. While some points from the earliest Tabun sample are retouched, a greater portion of Aduma points are retouched in every level. The pieces that are retouched at Aduma are more completely retouched (retouched in more areas of each piece) than the Tabun points. While inverse, or vental, retouch is rare throughout, it is actually most common in Tabun D. Bifacial retouch, on the other hand, is virtually absent in the Tabun sample and present at significant freuqencies throughout the Aduma sample, except for the uppermost level. Invasive retouch, a hallmark of the classic MSA, is found at low levels in Tabun D (alhtough not it B or C), but at Aduma, it rises along with the frequency of classic MSA points to a maximum in the Ardu II silt sites, then decreases slightly at the top.
Finally, striking platforms of complete pieces are virtually unmodified in the Tbaun assemblage but up to 50% are thinned or removed on the Aduma points. Overall, the earliest Aduma points from A-1 are most similar to the Tabun points from Level D. If one were to argue that the points for a moment of contact or expansion in either direction, this time period represents the most likely candidate. The later points in both areas dating to late OIS 5 and early OIS 4 are increasingly divergent in style (Brooks et al. 2005:248-249).
I quoted extensively to focus on the difference between the technologies and their different trajectories over time. Brooks and colleagues suggest that if the Tabun and Aduma samples were ever connected (informationally), then it was likely early -- before OIS 5. Through time, Tabun points remained relatively static in overall size but shortened, while Aduma (and ≠Gi) points got smaller and smaller.
Here is the concluding paragraph with respect to the projectile argument:
The Aduma points tend increasingly toward the dimensions of spear thrower darts or arrows, and hold point angle constant. Although ethnographic and archaeological examples of spear throwers are not known from any African site, spear throwers are present at a much later date on all the other inhabited continents and begin to appear at a time when African LSA armatures already fall within the size range of modern arrowheads (at <1.5 g). The early diminution of African stone armature may indicate that Africa passed through a spear thrower state at an earlier date (Brooks et al. 2005:251).
All of these arguments are functional rather than stylistic. The retouch observations fall into an interesting category, because these are (possibly) the strongest evidence concerning the attention of the point makers to their final form. Brooks and colleagues conclude that the functional stasis of the Tabun tools may be reflective of those aspects of early modern humans that inhibited their dispersal further into Eurasia.
Also, the paper includes some cool closeup photos of points to show impact damage on the tips.
Brooks AS, Yellen JE, Nevell L, Hartman G. 2005. Projectile technologies of the African MSA: Implications for modern human origins. In Hovers E, Kuhn SL, eds. Transitions Before the Transition: Evolution and Stability in the Middle Paleolithic and Middle Stone Age. Springer, New York, pp. 233-256.