Archaeology and Ethnoarchaeology of Mobility

Edired by Frederic Seller, Russell Greaves, and Pei-Lin Yu

University Press ofFiorida Gainesviiie/Taiiahassee/Tampa/Boca Raton Pensacoia/Oriando/Miami/]acksol1viiie/Ft. Myers '-Z-r) () G ogy, Harvard University, Cambridge, Mass. Kelly, R. L. 1992 Mobility/Sedentism: Concepts, Archaeological Measures, and Effects. Annudl From Atlatl to Bow and Arrow Review ofAnthropology 21: 43-66. Implicating Projectile Technology in Changing Systems Koster, H. A. 1977 The Ecology of Pastoralism in Relation to Changing Patterns of Land Use in the of Hunter-Gatherer Mobility Northeast Peloponnese. Ph.D. dissertation. University ofPennsylvania.

Moshkova, M. G. PEl-LIN YU 1992 Stepnaya polosa Asiatskoi chasti SSSR v skifo-sarmdtskoi vremi (Steppe Region of the Asiatic Part of the SSSR in the Scythe-Sarmanti Time). Nauka Izdatel'stvo, Moscow. Nance,]. D., and B. F. Bal1 In regional culture histories, the transition from large, broad-based projectile 1986 No Surprises?: The Reliability and Validity ofTest Pit Sampling. Americdn Antiq­ points to smaller, lightweight forms is often cited as a shift in the launching uity 51 (3): 457-483. method from spearthrower (or atlatl) to bow and arrow (cf. Aikens and Higuchi Plog, F. T. 1982: 109; Cabrera Valdez 1984: 279; Grayson 1994: 250). Much research on 1973 Diachronic Anthropology. In Resedrch and Theory in CurrentArchdeology, edited the projectile transition has focused on intrinsic attributes ofpoints in order to by C. L. Redman, pp. 181-198.]ohn Wiley, New York. separate them into discrete types (Bettinger and Eerkens 1999: 231; Beck 1998: Rosen, A. M., C. Chang, and F. P. Grigoriev 21). In addition to formal variation in stone points, the reduction method may 2000 Paleoenvironments and Economy of Iron Age Saka-Wusun Agro-pastoralists in differ. Analysis of one sample from northeastern North America showed that Southeastern Kazakhstan. Antiquity 74: 611-623. dart points are typically reduced from cores and arrow points from Bakes (Nas­ Rouse, I. saney and Pyle 1999: 251-252). 1972 Settlement Patterns in Archaeology. In lvfdn, Settlement, dnd Urbdnism, edited by Projectile point types are often used as chronological markers or "guide fos­ P.]. Ucko, R. Tringham, and G. W. Dimbleby, pp. 95-107. Duckworth, London. sils" (HuckellI996: 326), although variation in form may also result from me­ Shott, M.]. chanically conditioned behaviors, such as breakage, repair, and resharpening 1995 Reliability of Archaeological Records on Cultivated Surfaces: A Michigan Case (HuckellI996: 327). In a global survey, Pierre Cattelain (1997: 232) found that Study.Journal ofField Archaeology 22(4): 475-490. projectilepoint form alone is not correlated with hafting contexts or means of Stuiver, M., et al. launching. Charlotte Beck's (1998) analysis of examples from GatecliffShelter 1998 INTCAL98 Radiocarbon Age Calibration. Radiocarbon 40(3): 1041-1083. indicates that neck width is acted upon by selective forces and is useful in dis­ Wandsnider, L., and E. L. Camilli tinguishing darts from arrow points. Statistical tests for many archaeological 1992 The Character of Surface Archaeological Deposits and Its Influence on Survey sequences show that small, lightweight points replace or augment large, heavy, Accuracy.Journal ofField Archaeology 19(2): 169-188. Yablonsky, L. T. broad-based points (Shott 1997). 1995 The Material Culture of the Saka and Historical Reconstruction, Chapter 14. In The projectile transition occurred at different times, and at different rates, Nomads ofthe Eurasian Steppes in the Early Iron Age, edited by J. Davis-Kimball, throughout the world. The transition to bow and arrow never occurred in Aus­ V. A. Bashilov, and L. T. Yablonsky, pp. 201-240. Zinat Press, Berkeley. tralia. Atlatls and bows and arrows were used in tandem in the recent past in the Arctic, the North American Southeast, and parts of Mesoamerica. Recent efforts to explain the variation in scope and timing of the'projectile transition have focused on distinguishing between in situ development and diffusion, es­ pecially in North America (cf. Bettinger and Eerkens 1999; Nassaney and Pyle 1999), then proceeding to test models for different modes oftransmission. Rob­ 202 / Pei-Lin Yt. Projectile Technology in Changing Systems ofHunter-Gatherer Mobility / 203

..------~ ert L. Bettinger and Jelmer Eerkens (1999: 240) explicitly define the projectile 120 transition as a cultural evolutionary stage indicative ofa culture's versatility and innovativeness. 100 ~ Technology, however, unlike cultural markers such as style, articulates closely with adaptive strategies of resource use and mobility and therefore is expected 80 to co-vary with selective forces (Beck 1998: 23; Knecht 1997: 2). If selective c forces are less intense in a given case, technological behaviors may be adopted OJ () 60 and incorporated as part of the preexisting system. In this case the material re­ ill 0.. cord would show the new and old technologies coexisting, at least for a time. More intense selective forces and in situ development ofa new technology may 40 Large projectile point lead to total replacement of the old system and corresponding replacement in Dabsent the archaeological record. This chapter builds on the substantial body ofwork on projectile point classi­ .present fication and development ofregional chronologies by exploring the larger adap­ ':L ~unknown tive context of the projectile transition. In order to do this, it is necessary to A-e-lr ans ition fust-transition compare archaeological sequences that share similar projectile transitions but Site type are geographically distinct. Defining and explaining changes in a varied set of adaptive contexts should shed new light on the significance of the projectile Figure 9.1. Bar graph ofpre- and posr-transitional projectile points for the study areas. point transition for the larger issue ofcultural evolution. Lithic tool density (n tools per cubic meter of excavated sediment) in the 3 AREAS OF STUDY study areas tends to stay within a range of about 0-100 tools/m (Figure 9.2). Post-transition sites show a Y-shaped distribution after ca. 17,000 BP, in which The archaeological records of northeast Asia, southern Europe, and the North they become densely deposited or continue to be sparsely deposited. American Great Basin all indicate that large projectile points were either re­ Detailed information on sediment accumulation was not available for all placed by or augmented by smaller, lightweight points. sites, but light lithic tool deposition is apparently an important characteristic of The decrease in sites containing only large points is a defining characteristic of the projectile transition, regardless of time or place (Figure 9.1). Table 9.1. PrOjectile Point Transition Sites in the Analysis The following cultural sequences were selected because they include well­ Coastal Spain Japan North American Great Basin dated projectile point transition sites or levels within multicomponent sites: La Riera Suzuki Corn Creek Dunes (1) Late Paleolithic/Initial Jomon periods of central Japan and Hokkaido; (2) E[ Rascano Nogawa Danger Cave Late Solutrean/Early Magdalenian periods of Cantabrian Spain; and (3) Late EIJuyo Uenodaira Hogup Shelter Archaic/Early Prehistoric periods ofthe North American Great Basin. Table 9.1 Ekain Shirataki-Hattori dai Sudden Shelter shows the sites discussed in this chapter. Altamira Tachikawa Cowboy Cave Ambrosio Kidosaku E[ Castillo Horokazawa ESTABLISHING ARCHAEOLOGICAL RELATIONSHIPS E1 Pendo Midorigaoka Basic information on lithic tool types and counts, faunal species present, and Chunn Fukui methods ofexcavation is routinely recorded in study area archaeological site re­ Cuero de la Mina Natsushima Kamikuroiwa ports. These data are useful in determining the nature ofthe tool kit, the animals Kosegasawa targeted, and the depOSitional context ofartifacts. Study area sites or site levels Yubetsu-Ichikawa that bracket the proiectile transition were selected. 202 I Pei-Lin Yu ert L. Bettinger and Jelmer Eerkens (1999: 240) explicitly define the projectile transition as a cultural evolutionary stage indicative ofa culture's versatiliry and innovativeness. Technology, however, unlike.culturalmarkers such as style, articulates closely with adaptive strategies ofresource use and mobility and therefor~ is expected to co-vary with selective forces (Beck 1998: 23; Knecht 1997; 2). Ifselective forces are less intense in a given case, technological behaviors may be adopted. and incorporated as part ofthe preexisting system. In this case the material re­ cord would show the ne'Y and old technologies coexisting, at least for actime:-~ More .intense selective fo~~es and in situ devel0pnlent ofa new technology m~y' lead to total replacement ofthe old system and corresponding replacement in the archaeologiCal record. This chapt~r builds on the substantialbody ofwork on projectile point classi~ fication anddevelopment ofregional chronologies by exploring the larger adap~ tive conte~ ofthe projectile transition. In order to do this, it is necessaryt~i compare archaeological sequences that share similar projectile transitions b!:l:~ are geographically distinct. Defining and explaining changes in a varied setdfi . adaptive contexts should. shed new light on the significance of the projectije point transition for the larger issue ofcultural evolution. '

AREAS OF STUDY

The archaeological records ofnortheast Asia, southern Europe, and the-North ' American Great Basin all indicate that large projectile points were either re­ placed byor augmented ~y smaller, lightweight points. The decrease in sites containing only large points is a: defining characteristic ofthe projectile transition, regardless oftime or place (Figure 9.1). The following cultural sequences were selected because they include well­ dated projectile point transition sites or levels within multicomponent sites; (1) Late Paleolithic/Initial Jomon periods ofcentral Japan and Hokkaido; (1) Late Solutrean/Early MagdaIenian periods of Cantabrian Spain; and (3) Late Archaic/EarlyPrehistoricperiods ofthe NorthAmerican Great Basin. Table 9.1 shows the sites discussed in this chapter.

ESTABLISHING ARCHAEOLOGICAL RELATIONSHIPS

Basic information on lithic tool types and counts, faunal species present, and methods ofexca\i-afion is ro~tinely recorded in study are:a archaeological site re­ "'­ ports. These data are ust;ful in determiningthe nature ofthe tool kii, th~ animals t;trgeted, and the depositional context ofartifacts. Study area sites or site levels that bracket the projectile transition were selected. . . Projl(ctile Technology in Changing Systems ofHunter-GathererMobility / 203

120

100 100

80

....c Q) '-' 60 Q; a.

40 Large projectile point

Dabsent 20 .present

0 ~unknown Fl'"e-lransilion Fbst-trans ROO

Site type

Figure 9.1. Bar graph ofpre- and post-transitional projectile points for the study areas.

Lithic tool density (n tools per cubic meter of excavated sediment) in the study areas tends to stay within a range ofabout 0-100 tools/m3 (Figure 9.2). PosHransition sites show a Y-shaped di~tribution after ca. 17,000 BP, in which they become densely deposited or continue to be sparsely deposited. Detailed information on sediment ~ccurnulation was not available for all site;; but light lithic tool deposition is apparently an important characteristic of

Table 9.1. Projectile Point Transition Sites in the Analysis Coastal Spain Japan North American Great Basin La Riera Suzuki Corn Creek Dunes ElRascano Nogawa Danger Cave ElJuyo Uenodaira Hogup Shelter lllin Shirataki-Hattori dai Sudden Shelter Altamira Tachikawa Cowboy Cave Ambrosio Kidosaku El Castillo Horokazawa BPendo Midorigaoka Chufin Fukui Cueto de la Mina Natsushima Kamikuroiwa Kosegasawa Yubetsu-Ichikawa 204 / Pei-Lin Yu

1000 ... 800

600

(") E en (5 400 .9 z ... 200 o ... Transition 0 c9 ... 0 @J ,en> * ...... after,

-200 o before ~ -30 -20 -10 o

Earliest occupation date x 1000 B.P.

Figure 9.2. Transition by lithic tool density and 14e date. pre-transition sites in all regions, regardless ofage. The consistencyoflithic tooL density in pre-transition sites is remarkable, with a meanIof27.1 tools/m 3• Post? transition sites showa bimodal pattemoflight andheavydeposition, with·heaV;Y~' values averaging425.18 tools/m3• Post-transition sites also exhibit regional dq~J-;' tering, withdensesdithics in the Spanish Magdalenian cave sites, intermediate values inJapanese sites, and lowest values in the Great Basin. Lithic tool density may reflect the frequency or intensity ofoccupations or site formation processes such as differential sediment accumulation. H~terf gathererlandscape use is responsive to the timingandlocation ofresources (:Bin~ ford 1983). Iflithic tool density reflects huinan behavior rather than site for-ma­ tion processes, it is expected that tool densitywill correlate with archaeolog~cal traces ofsubsistence. In all three study areas the presence or absence ofanimal species is the most c9mmonly reported class of sub~istence data. Lithic tool density shows a 1;"7 shaped distribution. when plotted against the proportion oflarge to all terrestria.! mammal species (Figure 9.3). When the total assemblage ofterrestrial mammal species is composed ofabout 38% large animals, the Spanish cave sites ofEkiiri; El Juyo, and El Rascaiio exhibit a sudden increase in lithics density. Japanese Paleolithic sites are not represented due to poorbone preservation ofvolcanic soils. Great Basin sites show the lowest values for both lithic---- tool density and' presence oflarge mammal species. Projectile Technology in Changing Systems ofHunter-Gatherer Mobility / 205

1000

800

600

C") ~ 0 400 .9 z ~ 200 0 ~ q Transition 0 O~~T)()y ~O ~ after

-200 o before 0.0 .2 .4 .6 .8 1.0

N large sp.ltotal N. sp.

Figure 9.3. Transition by lithic tool density and large mammal species frequency.

In study sample sites the relationship between lithic tool class density and the frequency oflarge mammal species demonstrates that varying tool densities reRect real differences in site use rather than site formation processes. Ifthis is true, then pre-transition sites in all three study areas share low tool den.sities and therefordow discard rates despite variation in chronology, environment, and subsistence. Post-transition sites show a Y-shaped distribution in which Span­ ish ca"e sites diverge toward a new pattern ofhigh tool-discard rates. This pat­ tern suggests more frequent orlonger occupations. Great Basinsites consistently show low tool-discard rates and low representation,oflarge game, andJapanese sites are intermediate. Assemblage richness, or the number oftool classes, is divided by the number oftools to provide a measure for lithic tool class diversity. Inpre-transition sites

I lithic tool diversity stays within a range of0.01 to 0.03 regardless ofgqograpruc area or chronology (Figure 9.4), with one exception: the pre-transition level of Cowboy Cave in Utah. This site containedverylownumbers oftools but a mod­ erate number oftool classes, which increased the value ofthe ratio between the h two. Post-transition sites show a bimodal pattern in tool diversity as with tool density, in which diversity either rises quickly or stabilizes at the low, pre-transi­ . tion range. In sum, pre-transition sites or site levels show surprisingly consistent sparse lith,ktool density and low'tool class diversity. Post-transition sites show a lithic tool density threshold at a ratio of0.38 large mammals and at site dates of ca. 206 I Pei-Lin Yu

.12 o .. .10 .. .. en 0 E .08 z .. en Q) en en .06 ro <3 0 E .04 .. .. Z .. 00 0 .. 0 .. T~mo" .02 o 0 J.. after , CO 0.00 <9 o before ~ . . -30 -20 -10 0

Earliest occupation date x 1000 B.P.

Figure 9.4. Transition by n tool classes and 14C date.

17,000 BP. Two trajectories are visible, one showing a drastic increase in density' and the other maintaining stable pre-transition levels: Post-transition site tool class diversity shows a similar pattern. 'The studysample indicates that archaeological correlates ofthepre-transition stage ar~ consistentacross a broad range ofchronologies and environments. Low tool discard rates and lowtoolclass diversity are important definingcharaeteris­ tics ofpre-transition behavior that suggest shonCterm orinfrequent use ofsites and a conservative tool kit. \ In contrast, post-transition sites show significant divergence betweenr "ota" fashioned" low values for artifact discard and tool class diversity and new. very high values. The high values may indicate long-term or more frequent site use and a more diverse tool kit.

USING PROJECTILE WEAPONRY

These 'archaeological patterns suggest significant changes in site use and techno­ logical strategies during the projectile trap-sition. Projectile points may be used for many different purposes, but deSign constraints are generally imposed by their function as hunting tools (Greaves 1997: 313). Therefore'ethnographic in­ formation on hunting, practices is auseful framework for understanding----­ changes in site use and technology. l~rojectile Iechnology in Changing Systems ojHunter-Gathr:rerMobility / 207

Many authors work from the assumption that large lanceolate and small tri­ angular points respectively represent two specific hunting tools, the atlatl and the bow and arrow (Grayson 1994: 253; Straus 1983: 90; Sugihara 1973 in Aik­ ens and Higuclu':l982: 71). The best available comparative information on these two forms oflaunching Comes from experimental studies ofprojectile ballistics and from ethnographic hunting accounts.

Experimental Data: Range, Accuracy, and Momentum Experirrients in projectile technologyhighlight the contrasting abilities ofatlatls and bows and arrows. Cattdain (1997: 218) summarizes several experiments in central and northern Australia to arrive at a potential throwing distance for spearthrowers ofup to 180 m. Australian hunters' preferred thro~ing distances with a spearthrower, however, are between10 and 20 m (summarized in Cundy 1989: 17). Potential bow'and arrow throwing distance is about 100 m using traditional bows, but preferred throwing distances with bow and arrow are summar1zed by Cattelain (1997: 227) at about 9-25 m, beyond which accuracy is compro­ mised. Subtracting the actual from the potential throwing distance provides an estimate of the force differential between these weapons:_the spearthrower potential minus actual distance is ca. 160 m; for bow and arrow the difference is ca. 75 m or less. Therefore, holding range and projectile weight constant, the spearthrower delivers overtwice as much force upon impact, presuming equal weights for projectiles. This estimate can be refined by looking at the siniple momentum ofdifferent projectiles. Deceleration due to drag has been shown to be inversely proportional to the projectile's mass (Cundy 1989:34), so that high velocity increases the kinetic energy ofa projectile but also increases drag by an equivalent amount. MomentUm from heavy mass overcomes the drag factor; therefore heavier pro­ jectiles will penetrate farther than lighter ones when all other factors are held constant (Cundy 1989: 34; Dietrich 1996: 41). ~ Academics and modern bowhunters agree that mass is crucial to the con­ straints and opportunities of different projectiles. Although light weight and higher velocity produce ,a flatter trajectory and better accuracy at long range, greater mass delivers the greatest penetration and the best wounding potential (Dietrich 1996: 54). Launching velocity of ancient projectile points cannot be known for certain, but point weights can be compared to approximate mass. Great Basui dart shaft dimensions are known from archaeological specimens at Lake Winnemucca, Humboldt and Lovelock Caves, and Leonard Rockshelter (Spencer 1974: 41). An experimental untipped atlatl dart mainshafr based on those measurements had a length of 575 cm and weight of 67.5 gm (Spencer 208 / Pei-Lin Yu

1974: 41). No foreshaft was manufactured. The mean weight ofElko points from Danger Cave and Hogup Shelter (Aikens 1970: 46-47) is 4.4 gm. The total estimated weight ofa Great Basin dart-mainshatt assembly is therefore around 72gm; Dans from Cattelain's ethnographic sample (1997: 229) fall into two size classes; between 30 and 63 cm in length and 55-200 gill in weight for Arctic examples and between 190 and 460 cm in length and 250-600 gm in weight for Australian examples. Arrows are divided similarly: in open temperate or forested tropical areas they are 43-110 cm long and weigh 15-40 gm. Arrows used in open tropical (usually savanna) areas are 110~21O em long and weigh 35-88 gm (Cattelain 1997: 229). Among smaller projectiles, arrows are 0.55 the weight ofadad darts; among larger projectiles, arrows are 0.15 the weight ofadad darts. The average ofthese values is a 0.35 ratio for arrow to adad dart weights. The relationship between the masses of the two projectiles can be approximated by the relationship between theirweights. Given that momentum = mass x velocity, the momentum ofdarts is from two to seven times greaterthan arrows, with an average ofabout3.5 times greater, ifvelocity is held constant. This ratio clearly affec:ts the size ofintendedprey. An adad dart delivers a more forceful blow and penetrates deeper. The point must be correspondingly robust to withstand high impact. Large animals have thicker hides, more meat ~ass, and· denser bones, all ofwhich may deflect orbreak the point. In order to maximize lethal effect, it is important to decrease the drag ofthe projectile head once it has struck so that it penetrates deeper (Cundy 1989: 34). Two characteristics can do this: a very sharp wtting edge and a broad head thai accommodates the shaft as , it follows (Cundy 1989: 34). /:,1 'it. Litde enetgy is needed once the skin is penetrated, however, and large animals " can be killed effec:tively by small arrows (Cundy 1989: 36; Knight 1975: 253 in ;1 Cundy 1989). Ifadads and arrows are equallylethal, whydid some hunters retain adads long after bows and arrows came into use?

Ethnographic Data: Contexts ofUse

Ethnographic Studies s~ow that the spearthrower is used almost exclusively in open environments (Cattelain 1997: 219). At the time ofEuropean exploration, atlads were commonly used in only two regions of the world: desert Australia and the Arctic (Cattelain 1997: 215). Adads are shockweapons (Churchilll993: 18; Hutchings 1998: 13) that rapidly debilitate prey and reduce the need for tracking. Australian hunters have edmpletely_impaled l~ge game such as kan­ garoos with adad spears (Tindale 1925). Large arrows in open tropical settings work the same way. The Pume ofVenezuela use long arrows whose weight adds Projectile Technology in Changing Systems ofHunter-Gatherer Mobility / 209 , shock to impact, and length inhibits movement after the animal is hit. This is .particularly important in prey of high escape risk, such as arboreal, flying, or swimming animals, (Oswalt 1976: 81). For example, coastal Eskimo hunters use the bow and arrow for terrestrial game and the atlatl for marine mammals and aquatic birds (Cattelain 1997: 215). In contrast, bows are used in every environment and exhibit great variation in form and mechanics. Small arrows work as delivery devices for small, sharp points and are often armed with lethal substances in the case oflarge prey (for example, by desert hunters ofSouth Mrica) or arboreal prey (as in the case ofthe South American tropics). Bows and arrows offer a flexibilityof use contexts that cannot be matched by the atlatl. Atlatl spears are bulkier than arrows: 1.3 to 3 times longer and 2 to 5 times heavier. Norman B. Tindale (1925: 98) noted that Australian hunters prefer bamboo for spear shafts ifavailable, in order to increase the number ofmissiles that they can carry. Ahunter can carry fewer spears than arrows, all other things beingequal. In addition, thespearthrower itselfis bulkier, and often heavier, than a typical bow (Cattelain 1997: 217). Hunters who. use atlatls or long arrows must carry their projectiles by hand, which limits the number that can be carried on trips to four to s~ven (Bartram 1997: 335; Greayes, personal communication, 1999). Up to twentyarrows might be carried in a quiver, however (Hitchcock and Bleed 1997: 348). More projec­ tiles allow for more shots per trip and longer search times. Search time duration is directly related to potential encounters with resources on hunting trips (Greaves 1997: 310). Also, less time is lost in resharpening or other repair, which may not be necessary ifenough arrows are carried. Modern bowhunting and archaeological experiments indicate that smaller project4es are flatter in trajectory and more accurate on target (Cattelain 1997: 230;° Churchill 1993: 18; Dietrich 1996:54). Thus bows and arrows, while still lethal for larger game, allow for· smaller target sizes if the range is held constant (Churchill 1993: 20). Finally, the atlatl requires space for the throw and preferably space in front to step forward. Although Eskimo atlatls are small and may be launched from a seated position in a kayak (Nelson 1899: 151-152), the normal position among Australian hunters stalking terrestrial game is standing and walking (Cattelain 1997: 219). The bow and arrow are held close to the body and may be launched when standing still or seated, requiring less space (Figure 9.5). Hunters attetp.pting to narrow the distance between themselves and preyhave two options: concealment and disguise (Dietrich 1996: 160). Disguise involves carrying extra gear, but concealment requires only the presence ofminimal ob­ stacles in the prey's line ofsight. In a comparison between large and small bows and arrows, Lawrence E. Bartram (1997: 340) found that large arrows hamper 210 I Pei-Lin Yu

, .. ~-"-- ...... I' - ..... I' \, I \ / \ I I \ I \ I \ I \ I ~ I

Figure 9.5. Space requirements for launching different projectiles. the hunter's ability to stalk by crawling or shoot while hidden in vegetative cover. Small bows and arrows permit the hunter to fire onpreywhilein heavyvegetative coveror to crawl closer ifcover is sparse. To summarize, ethnographic accounts show that adads immobilize prey upon impact and thus reduce the risk ofescape. This is particularly important with large, fast animals or animals that can escape into environments beyond the hunter's reach, such as air and water. At some point, these benefits became out­ weighed by the arrow's light weight, flatter trajectory, and versatility in launch­ ing. Smaller missiles such as bow and arrows offer the hunter longer search times and more shots per trip. Smaller prey can be targeted successfully, and stalking is possible in dense or ground-level cover. The hunter might spend less tirnein repair or maintenance when equipped with many projectiles. Extended trip times, higher number ofshots, shrinking prey size, and exploi­ tation ofpreviously unused habitats suggest the hunter's need to maximize the number ofsuccessful,shotsper trip. This in turn indicates widening diet breadth and the need to decrea~epursuittime'averaged across a growing llumber·ofprey classes (Churchill 1993: 21). Pre-transitionsites and post-transition sites that stay atpre-transition values are characterized byselective use ofthe landscape andless 1 need to hunb in heavily,vegetated areas. For example; Great Basin hunters may nothave needed toshootwhilestanding!in densevegetatiou but found ,the bow and arrow usefulfor targeting the small game characteristic ofthe desert.

CLIMATIC CONDITIONERS OF THE TRANSITION

The transition to bow and arrow occurred at different times and at differenuates in the three study areas, but both Japanese and Spanish transitions predate the Great Basin. Why? Use coutens of the bow and arrow suggest that the answer lies in the hunter's choice to include heavily vegetated terrain in the normal sub­ sistence round. An early projectile transition predicates a high percentage of heavily forested habitat in the general environment. , , Rrojectile Technology in Changing Systems ofH.unter-GathererMobility / 211 \

Temperature and Growing Seasons As mentioned above, small projectile points appear early in Cantabrian Spain and Japan (ca. 1),000-14,000 BP and ca. 14,000-12,000 BP, respectively). The transition is distinct and is used in both regions as a chronological marker. In the Great Basin the transition took place from ca. 5000 to 1700 BP and was so gradual that the dates are still subject to debate. Effective temperature (ET: average of temperatures taken at the beginning and end ofthe growing season) in the three study regions has varying effects on the transition (values taken from Binford's database; also see Binford 2001: 258). The "bow and arrow" line can be tracedalongeffective temperature values, show­ ing the latest transition in the Great Basin at a value ofabout 13.5" C in the left of the figure (Figure 9.6). Japan's effective temperatures fall on both sides ofthe Great Basin, and Spain has the highest effective temperature values on the right ofthe figure. The bow and arrow line suggests that the transition is delayed at ET values ofabout 13° C.

Precipitation and Vegetative Cover Vegetative growth is regulated by the interaction between temperature and rain­ fall. Figure 9.'ishows how rainfall conditions the net above-ground productiv­ ity, or grams ofplant material added per m2 per year, in each study area (values are fr~m Binford's database). Overall, higher rainfall results in higher vegetative growth ra~es. SouthernJapan and Spall have higher values, northernJapan is in­ termediate, and the Great Basin shows very low ra~fall and vegetative growth.

0 ... a.: IIi ... a C:> -10 ... 0 0 ...... ~ CO x ... 2 0 lU u '0 0 c -20 0 <:e> ~ 0­ :::l 0 0 '0 >. -30 L: lU 0 Transition W ... after

-40 o before 12.0 12.5 13.0 13.5 14,0 14,5 15.0 15,5 16,0

ET (effective temperature)

Figure 9.6. Transition by effective temperature and 14C date. 212 / Pei-Lin Yu

3000TI------, Southern Japan

E E .E: 2000 ~ T t: .~ Coastal Spain ~ "iii o TT ::J t: t: TT ~ 1000 Northern Japan T ~ Region () -~~ @ Great Basin o Spain @~ . I • T Japan oIii , , ii' o 200 400 600 800 1000 1200 1400 1600

NAGP (plant productivity)

Figure 9.7. Study regions by annual rainfall and net above-ground productivity.

Earliest transitions occur in Spanish and southern Japanese sites, which both are characterized by high annual rainfall, high net above-ground productivity, and high effective temperatures. The Great Basin sites, with lowest annual rain­ fall, low net above-ground productivity, and moderate effective temper~tures, consistendy show/ the latest transition. Northern Japan and Hokkaido, with coldest temperatures in the sample and low to moderate rainfall, are intermedi­ ate to Spain and the Great Basin. These results indicate that high aridity and cold temperatures delay the transition but that aridity is a more significant deterrent. The Great Basin's position along the bow and arrow line strongly indicates a.different selective context from Cantabrian Spain and Japan. Recent research indicates that the transition in the Great Basin may have been a historical event ofadoption rather than an adaptive process ofinnovation (Blitz 1988). If.this is the case, then selective forces were not intense enough to encourage projectile in­ novation; high aridity and low plant biomass are associated with low population density and high mobility'among hunter-gatherers (Binford 1983). The blurred Great Basin transition across several thousand years and significant geographic variation indicates that the selective forces favoring bow and arzow technology were in play, but at a lower degree ofintensity. R~ojectile Technology in Changing Systems oflfunter-GathererMobility / 213

Paleoclimatic Triggers Ifpaleoenvironments affected the chronology of the projectile transition, cli­ matic changes'in-the study areas may have provided discernible trigger events. InJapan the projectile point transition occurred area. 14,000 13F in the south and moved north over time (Aikens and Higuchi 1982: 87). The terminal glacial climate ~f the late Paleolithic and earlyJomon was several degrees cooler than at present (Pearson 1977: 1239). Pollen records indicate that Hokkaido was open parkland with spruce, beech, and elm and that the maih body ofJapan was thick­ ly forested with oak, elm, and beech. In the early Holocene these forests began to shift to 'broadleaf evergreen forests, and climatic warming iriitiated mirine transgression ofthe Japanese coastline (Pearson 1977: 1240). This transgressioii probably reached its maximum during the earlyJomon, ca. 9000 BP (Pearson 1977: 1240), swamping the coast, moving into topographically varied areas, and offermg a greater diversity ofcoastal swamps, bays, and estuaries than before (Ai­ kens et al. 1986: 16). The$panish Solutrean period overlaps with a glacial maximum (ca. 20,000­ 17,000 BP) and shifts to the Magdalenian period during inters~adial warming (Straus 1991: 89). Warming glacial climates in the open coastal plains along the coast ofCantabrian Spain allowed small thickets oftrees to devel6pinwind-shel­ tered areas (Straus and Clark 1986: 1.36). Marine transgression had begun by the early Magdalenian, but little coastline was lost (Straus and Clark 1986: 136). The projectile point transition in the Great Basin occurred muchlater than in the other two areas, in the late Holocene. Duringthis time the desert underwent coolingtemperatures and increasingmoisture from an aridity maximum at about 7000-6000 BP. Pollen data indicate that juniper (which follows moisture) be­ came more frequent and that pilion pine moved into the Great Basin from the south (Grayson 1994: 221-222). Overall, cooling temperatures and moisture in the eastern Great Basin caused grass to replace sage and scrub in open areas and. allowed the treeline to move down in elevation (Grayson 1994: 221-222). In all three study areas an environmental rebound from climatic extremes of either Pleistocene cold or mid-Holocene aridity accompanies the projectile point trari'!lition. InJapan and Spain, which were more denselyvegetated to begin with, the transition occurred earliest. In the desert ofthe Great Basin post-Pleistocene warming did not constitute a rebound and in fact led to the prolonged drought ofthe middle Holocene. Only in the late Holocene did the Great Basin return to early Holocene vegetative productivity (Grayson 1994). Through analysis ofthe three studyareas this chapter has restated relationships that are well known to human ecolOgists: climate conditions landsCape charac­ teristics such as resource distribution, and human hunters adjust their mobility to resource availability in time and space. Technology reflects these adjustments. 214- / Pei-Lin Yu

Although projectile technology is observed to change at different times and at different rates, these differences may be quantitative rather than qualitative, simi­ lar processes playing out indifferent theaters (c£ Binford 2001: 33). Another important factor conditions hunter mobility and has the potential to affect the timing oftechnological shifts: the proximity ofother hunters. This brings the analysis full circle. The archaeological values oflirhic tool density and assemblage richness, and ethnographically known use contexts ofhunting gear, can be linked with envi~~nmentalcharacte~istics,using human demographics.

INFERRING POPULATION GROWTH AND REORGANIZATION OF HUNTING TEltRITORIES

One class ofpost-transitional sites in the study areas shows evidence of redun­ dant, frequent site use (high lithic tool density tied to discard) and generalized subsistence (high assemblage richness tied to diverse processing needs). The bow and arrow have been demonstrated to suit longer trip times, more shots per trip, and targetillg ofsmall prey in previously unfavorable forested terrain. All other things beingequal, these adaptations are consistentwith decreases in territory for aforaging group (Binford 1983: 211; Kelly 1995: 151). , Ifhunter-gatherers use mobility as a security option to be abandoned only under duress (Binford 1983: 210), then the factor most likely to decrease resi­ dential mobility is the proximity of neighbors. This should also hold true for hunting mobility. Projected hunter-gatherer population densities were derived for thestudysites based on ethnographicallyknown hunter-gatherer densities in similar environments (values from Binford 2001). These values show that high population density would have been reached first in Spain, followe& by north­ ern Japan, southern Japan, and finally the Great Basin, which patterns well with known dates ofthe projectile point'transition (Figure 9.8). In.arid environments where territories are large and mobility high, packing thresholds are not expected to be reachedveryquickly. But changing pop~ation distributions in highly productive environments may have fuel~d early projectile transitions. In areas whClTe, preferred hunting niches filled and hunring territory overlap increased, huntelTs could maximize hunting returns in 'shrinking terri­ tories by induding more stalking habitats and diverse prey sizes. This favored a lightweight projectile offlexible use context and high accuracy. Features of the landscape, including hunting territories and associated residential si~es" began to he used more often or fot; longer periods. Diversification of tool classes sug­ gests a shift from procurement to processing; various scra~rs and awls for hide working constitute most of the assemblage richness in Spain and ground stone and crescent chipped stone forms for vegetable processing in Great Basin. Such

\ \ Pl;ojectile Technology in Changing Systems o/Hunter-GathererMobility / 215 \

Figure 9.8. Regional variation in projected forager populatio~ density by 14C date. tactical shifts do not necessarilysignify an absolute decrease in mobility, but they do reflect strategic adjustments in scope and scheduling as a response to gradual packing ofthe landscape.

Changing Modes ofWarfare Changes in hunting strategies and technology certainly affected the conduct of warfare (Nassaney and Pyle 1999: 258). The results presented in this chapter suggest that arrows allowed the development ofambush or guerrilla tactics. This kind ofwarfare, which relies on stealth and flexibility, is less formalized and more lethal than melee-style warf~rc of atlatl users such as those in desert Australia (Gason 1879). Arrows as weapons may also have permitted longer war forays and added the potential to launch more projectiles per foray.

CONCLUSION

Changes in tac.tics or behavior reflect adaptive responses to certain conditions, which the archaeologist may discern through the application ofrelevant bodies ofinformation to archaeological data. The projectile rransition represents a tacti­ cal adjustment to changing resource availability: the addition ofa new hunting niche. The timing and the rate of this process are predictably associated with 216 I' Pei-Lin Yu the intensity of selective forces acting upon human populations; factors such as population density, mobility, and resource availability are all conditioned by local environments. In conclusion, defining the projectile transition (and other changes in type and frequency of materials associated with mobility, technology, ane!- subsis­ tence) iIi strictly evolutionary terms! may not be warranted. Making' a strategic adjustment to changing conditions, or choosing not to do so, does not in itself constitute an evolutionary threshold. lh.e true Rubicon, the ability to adopt or develop new strategies of mobility and resource use and to exploit those strate­ gies situationally, was crossed long before the projectile transition took place. ­

REFERENCES

Aikens, C. M. 1970 Hogup Cave. University of Utah Anthropological Papers 93. University of Utah Press, Salt Lake City. Aikens, C. M., K. M. Ames, and D. Sanger 1986 Affluent Collectors at the Edges ofEurasia and North America: Some Compari­ sons and Observations on the Evolution ofSociety among Northern-Temperate Coastal Hunter- Gatherers. In PrehistoricHunter-Gatherers in]apan, edited by T. Akazawa and C. M. Aikens, pp. 3-26. University ofTokyo Press, Tokyo. Aikens, C. M., and T. Higuchi 1982 Prehistory of]apan. Academic Press, New York. Altuna,]., and]. M. Merino 1980 Elyacimientoprehistdrico de la Cueva de Ekain, Deba, Guipuzcoa (The Prehistor~c Site ofEkain Cave, Deba, Guipuzcoa). Sociedad de Estudios Vascos, Guipuzcoa, Spain. Altuna,]., and L. G. Straus 1976 The.Solutrean of Altamira: The Artifactual and Faunal Evidence. Zephyrus 26­ 27: 175-182. Barandiaran, 1., L. G. Freem~,]. Gonzalez-Echegeray, and R. G. Klein 1985 Excavaciones en la Cueva del]uyo (Excavations in El]uyo Cave). Ministerio de Bellas Artes y Archivos, Madrid. Bartram, L. E. 1997 A Comparison ofKua (Botswana) and Hadza Bow and Arrow Hunting. In Pro­ jectile Technology, edited by H. Knecht, pp. 321-344. Plenum Press, New York. Beck, C. 1998 Projectile Point Types as Valid Chronological Units:In Unit Issues mArchaeology: Measuring Time, Space, and Material, edited by A. Ramenofsky and A. Steffen, pp. 21-40. University ofUtah Press, Salt Lake City. , Projectile Technology in Changing Systems ofHunter-Gatherer Mobility / 217

Bettinger, R. L., and J. Eerkens 1999 Point Typologies, Cultural Transmission, and the Spread of Bow and Arrow Technology in the Prehistoric Great Basin. American Antiquity 64(2): 231-242. Binford, L.R. '---­ 1983 In Pursuit ofthe Past. Thames and Hudson, New York.­ 2001 ConstructingFrames ofReference. University ofCalifornia Press, Berkeley. Blitz,J. H. 1988 Adoption ofthe Bow and Arrow in Prehistoric North America. North American Archaeologist 9: 123-145. Cabrera Valdez, V. 1975 El yacimiento Solutrense de Cueva Chufin (Riclones, Santander).XIVCongreso Nacional de Arqueologia: 157-164. 1984 El yacimiento de la Cueva de "El Castillo" (Puente Viesgo, Santander). Biblioteca Praehistorica Hispana, Madrid. Cattelain, P; 1997 Hunting during the Upper Paleolithic: Bow, Spearthrower; or Both? In Projectile Technology, edited by H. Knecht, pp. 213-240. Plenum Press, New York. Churchill, S. 1993 Weapon Technology, Prey Size Selection, and Hunting Methods in Modern Hunter-Gatherers: Implications for Hunting in the Paleolithic. In Hunting and Animal Exploitation in the Later Paleolithic and Mesolithic ofEurasia, edited by G. Peterkin, H. Bricker, and P. Mellars, pp. 11;-24. Archeological Papers of the American Anthropological Association 4. Washington, D.C. Conde de la Vega del Sella 1916 Paleolitico de Cueto de la Mina (Asturias). Comisi6n de Investigaciones Paleon­ to16gicas y Prehist6ricas, Memoria No. 13. Museo Nacional de Ciencias Natura­ les, Madrid. Corch6n,M. 1971 Materiales solutrenses de la Cueva Santanderina de El Pendo. Zephyrus 21(11): 7-21. Cundy,B.]. 1989 Formal Variation in Australian Spear and Spearthrower Technology. BAR Series 546. BAR, Oxford. Dietrich, D. 1996 BowhuntingBig Game. North American Bowhunting Association, Chicago. Echegaray,J. G., and I. Barandiaran-Maestu 1981 ElPaleolitico Superior de la Cueva delRascafio (Santander). Ministerio de Cultura, Santander. Esaka, T., aiid S. Nishida 1967 Aichi~ken Kamikuroiwa iwakagi. InNihon no doketsu iseki, edited by Nihon Kok" ogaku kyokai koketsu iskei chosa tokubetsuIinkai, pp. 224-236. Heibonsha, To­ kyo. 218 I Pei-Lin Yu

Gason, S. 1879 The Manners and Customs ofthe Dieyerie Tribe ofAustralian Aborigines. In The Native Tribes ofSouth Australia, edited by]. D. Woods, pp. 253-307. E. S. Wigg and Son, Adelaide. Grayson, D. K. 1994 The Desert's Past. Smithsonian Institution Press, Washington, D.C. Greaves, Russell D. 1997 Hunting and Multifunctional Use of Bows and Arrows: Ethnoarchaeology of Technological Organization among the Pume ofVene.-:uela.InProjectile Technol­ ogy, edited by H. Knecht, pp. 287-320. Plenum Press, New York. Hayashi,K. 1968 The Fukui Microblade Technology and Its Relationships in Northeast Asia and North America. ArcticAnthropology 5(1): 128-190. Hitchcock, R., and P. Bleed 1997 Each According to Need and Fashion: Spear and Arrow Use among San Hunt­ ers ofthe Kalahari. In Projectile TecJ;nology, edited by H. Knecht, pp. 345-370. Plenum Press, New York. Huckell, B. B. 1996 The Archaic Prehistory ofthe North American Southwest.Journal ofWOrld Pre­ history 10(3): 305-373. Hutchings, Karl w: 1998 The Identification ofPaleoindian Fluted Point Delivery Technology through the Analysis ofLithic Fracture Velocity. Paper presented at the 63rd Annual Society for American Archaeology Meetings, Seattle. Jennings,]. D. . 1980 Cowboy Cave. UniversityofUtah Anthropological Papers 104. University ofUtah Press, Salt Lake City. 1957 Danger Cave. University ofUtah Press, Salt Lake City. Jennings,]. D., A. R.Schroedl, andR. N. Holmer 1980 Sudden Shelter. University of Utah Anthropological Papers 103. University of Utah Press, Salt Lake City. Kato; S., andH. Oi 1961 The Midorigaoka Site, Kunneppu, Tokoro County, Hokkaido.Minzokugaku Ken­ kyu(Tokyo) 26(1): 24-30. In Japanese with English summary. Kellar,].H. 1955 The At!at!in North America. Prehistory Research Series 3. Indiana Historical So­ ciety,Indianapolis. Kelly, R. L. 1995 The Foraging Spectrum. Smithsonian Institution Press, Washington, D.C. Knecht, Heidi -----... 1997 Introduction. In Projectile Technology, edited by H. Knecht, pp. 1-3. Plemun Press, New York. Projectile Technology in Changing Systems ofHunter-Gatherer Mobility / 219

Knight,B. 1975 The Dynamics ofStab Wounds. Forensic Science 6: 249-255. Nakamura, K. 1960 Kosegasawa doketsu. Nagaqka shiritsu kagaku hakubutsukan Kenkyu chosa Ho­ koku, Vol. 3. Translated in Aikens and Higuchi 1982. Nassaney, M. S., and K. Pyle 1999 The Adoption of the Bow and Arrow in Eastern North America: A View from Central Arkansas. American Antiquity 64(2): 243-263. Nelson,E. W. 1899 The Eskimo about the Bering Strait. Eighteenth Annual Report of the Bureau of American Ethnology to the Secretary ofthe Smithsonian Institution, 1896-1897­ Smithsonian Institution, Washington, nc. Oswalt, W. H. 1976 An Anthropological Analysis ofFood-Getting Technology. John Wiley and Sons, New York. Pearson, R. 1977 Paleoenvironment and Human Settlement inJapan and Korea. Science 197/4310: 1239-1245. Ripoll-Lopez, S. 1988 La Cueva de Ambrosio, Almeria, Spain. BAR International Series S462(i). BAR, Oxford. Serizawa, c., and Y. Kamaki 1967 Fukui Cave, Nagasaki Prefecture. In Nihon no doketsu iseki, edited bySelect Com­ mittee for the Investigations of Cave Sites. Select Committee for the Investiga­ tions of Cave Sites, Tokyo. InJapanese with English summary. Shirataki Research Group 1967 Shirataki iseki no kenkyu. University ofHokkaido, Sapporo. InJapanese with Eng­ lish summary. Shott,M. 1993 Spears, Darts, and Arrows: Late Woodland Hunting Techniques in the Ohio Val­ ley. American Antiquity 58: 425-443. 1997 Stones and Shafts Redux: The Metric Discrimination ofChipped Stone Dart and Arrow Points. American Antiquity 62: 86-101. Spencer,L. 1974 Replicative Experiments in the Manufacture and Use of a Great Basin Atlatl. In GreatBasin AtlatlStudies, edited by T. R. Hester, M. P. Mildner, and L. Spencer, pp. 37-60. Ballena Press, Ramona, Cali£ Stra~,L. G. 1983 From Mousterian to Magdalenian: Cultural Evolution Viewed from Vasco­ Cantabrian Spain and Pyrenean France. In The Mousterian Legacy, edited by E. Trinkaus, pp. 73-111. BAR International Series 164. BAR, Oxford. 220 / Pei-Lin Yu

1991 Epipaleolithic and Mesolithic Adaptations in Cantabrian Spain and Pyrenean France.JournalofWorld Prehistory 5(1): 83-104. Straus, L. G., and G. A. Clark 1986 La Riera Cave. Anthropological Research Papers No. 36. Arizona State Univer­ sity, Tempe. Sugihara, S. 1973 Point TOol Culture ofUenodaira, Nagano Prefecture, Japan. Meiji University, To­ kyo. InJapanese with English summary. Sugihara, S., and C. Serizawa 1957 Shell Mounds of the Earliest Jomon Culture at Natsushima, Kanagawa Prefec­ ture, Japan. Memoirs ofthe TOkyo Archaeological Society 3(2): 1-33. In Japanese with English summary. Sugihara, S., and M. Tozawa 1975 Microlithic Culture ofShirataki-Hattoridai. Meiji U~iversity, Tokyo. In Japanese with English summary. Suzuki,M. 1974 Chronology ofPrehistoric HumanActivity inKanto,Japan, Part II: Time-Space AnalysiS of Obsidian Transport. Journal ofthe Faculty ofScience (University of Tokyo) section 5: anthropology 4, no. 4: 396-469. Thomas, D. H. -, . 1978 Arrowheads and Atlacl Darts: How the Stones Got the Shaft. American Antiquity 43: 461-472­ Tindale, N. B. 1925 Natives ofGroote Eylandt. Records ofthe South Australian Museum 3: 61-134. Williams, P. A., and R. I. OrIffis 1963 1he Corn Creek Dunes Site: A Dated Surftce Site in Southern Nevada. Nevada State Museum Anthropological Papers No. 10. Nevada State Museum, Carson City. Yoshizaki, M., S. Segawa, and M. Ishikawa 1960 Hakodate: Hokkaido shiritsu hakodate hakubutsukan (no English title or page numbers available). Kiyo 6. InJapanese with English summary. Yubetsu-Ichikawa Research Group 1982 Yubetsu-Ichikawa Iseki (The Yubetsu-Ichikawa Site). Yubetsu-cho kyoiku Iinkai, Yubetsu, Japan. InJapanese, translated in Aikens and Higuchi 1982.

'­