Atl atl replicate study

By Jim Dunbar


This study was undertaken to test the assumption that the Florida assemblage of carved mammoth ivory artifacts were utilized as spear foreshafts. Dr. S. David Webb and I (submitted for publication) contend the Florida specimens were used as foreshafts rather than for some other function, such as osseous (bone, antler or ivory) projectile points as was the case in the European Upper Paleolithic.

Webb and I found support for the foreshaft hypothesis based on the frequency and location of green bone fractures on the Florida sample. Green bone fractures or breaks can be differentiated from breaks which happen after a few seasons of weathering when the osseous (bone, brittle antler and ivory) material relinquishes its elastic properties and becomes brittle. Of 39 measurable specimens 51% of the sample displayed green bone fractures. Green bone fractures were most common on the beveled end accounting for 76.2% of the green fractures. Fractures on the beveled platform displayed a jagged, irregular tear indicative of compression fracturing. That is where we reckon that the base of a stone spear point had been hafted to the foreshaft, causing breakage.

However, in developing the data on the ivory artifacts, Webb invited Heidi Knecht to inspect the Florida specimens. Knecht, who had done her doctoral research on European Upper Paleolithic osseous points, cautioned Webb that beveled end breaks similar to the breaks on the ivory shafts also occur when they are used as projectiles. In her dissertation, Knecht discussed Gravettian single-bevel bone points, their specifications, manufacture and use. In another study Dave Guthrie also found that breaks along single beveled bone shafts used as projectiles occurred at both the pointed and beveled ends. Because of the caution expressed by Dr. Knecht and the contradictory findings of Guthrie, an experiment using ivory and other materials for foreshaft material was deemed necessary. Some preliminary results are presented here. The idea was to propel spears with attached foreshafts using a spearthrower.



Two spearthrowers or atl atls were used to propel the foreshaft-tipped spear, one with two finger notches and the other with a straight hand grip. The finger notched or forward-grip atl atl is based on a modified Key Marco design. The straight handled or side-grip atl atl is based on a replica of a European Upper Paleolithic design. The forward-grip atl atl was constructed of pine wood and weighed 224 grams. It had a central groove running down its length for a shaft rest. The groove also allowed the proximal spear socket to be run up to the distal end of the atl atl where the spear socketed into a small upturned antler spur. This design proved to be easier to load and maintained the spear in throwing position. The side-grip atl atl was constructed of a composite of cypress wood and deer antler. The distal 1/5th is constructed from a piece of curved antler having a small hook at the end. The wood and antler sections were fastened by dowel-pin and socket joint and cemented with hide glue. This joint was bound with false sinew and hide glue to provide strength. This atl atl weighed 142 grams. Its design requires the user to visually observe the loading of the spear socket into the atl atl spur. It also requires the user to hold the spear in place in a somewhat awkward position, tightly pinched between the thumb and forefinger, before each throw. If the user is not attentive, the spear socket will dislodge from the atl atl spur requiring the user to reload before making a throw. This spear thrower has no shaft rest to minimize spear dislodging. In this experiment the forward-grip atl atl spanned a length of 38 cm from its distal hand grip to its spur, while the side-grip atl atl extended 45.7 cm between corresponding points.


A functional antler socket was made from a section of elk antler tine. This socket is 7.51 cm long with a conical socket 5.58 cm deep. It is also drilled on the opposite end with a straightsided hole 1.93 cm deep. It was seated onto the wooden spear shaft with dowel pin and socket joint and required no glue or hafting due to its tight fit. The conical hole was bored into the antler with a steel reamer which has the same angle of taper as the mean of the data from our ivory foreshaft study. Several wooden spear shafts were handmade but only the best one was used with a foreshaft attached. Many thanks are extended to Captain Way Grisset of the Tallahassee Fire Department who also owns and operates a part-time saw mill. He not only furnished the wooden shaft material, but also produced the first shafts and lessons on how to hand whittle square stock into round dowels using a block-plane. Shafts of oak, cypress and pine were made, but the oak shaft was fixed with the antler socket and used to carry the spear foreshafts to their targets. This shaft is 1.778 m long by 1.5 cm in diameter. The fletching starts about 5.5 cm below the proximal end of the spear shaft so that the feathers do not get damaged by the atl atl. Ken Home of the Museum of Florida History fletched the shaft in a Cherokee style with turkey feathers. The completed spear shaft used for testing foreshafts weighs 227 grams (with fletching, sinew binding and antler socket without foreshaft).


Many thanks to Claude Van Order who graciously demonstrated his considerable stone flaking (flint knapping) and bone carving skills. We met Claude at the Ginnie Springs youth camp jamboree where he was demonstrating flint knapping, and Jimmy Sawgrass spoke of the legends and lifeways of the 19th century Seminole indians. Wayne Grissett, Joe Latvis, Rob Patton, Jerry Gramig, Chris Lewis, Dean Quigley and Mark Muniz were among the group in attendance.

Silicified coral was prized by Archaic indians who routinely fashioned spear points from it about 5,000 years ago. In Archaic times it was heat-treated before it was flaked into final shape. Claude, however, made a fluted Clovis point from unaltered silicified (agatized) coral (see fig. 1). His tools consisted of stone hard hammers, bone and antler soft hammers, and an abrader rock to grind the flake extraction platforms.

Late that evening, at the Steamboat Hotel in downtown Branford, Claude replicated the manufacture of an ivory foreshaft. I had already supplied him with a template of average dimensions for constructing a foreshaft. I had already constructed one of oak wood using modern tools (a bench sander and various saws). It was Claude's task to make one the old fashioned way.

Dr. Webb furnished a piece of ivory cut from the proximal end of an African elephant tusk which was 17 cm in length. The 17 cm length is actually short for foreshaft production but slats of that approximate length had been sent to Claude who had pretreated them by soaking in water for about three days.

While Joe Latvis videotaped the sequence, Claude manufactured a replica foreshaft from a slat in about 20 minutes. He used a 454 gram uniface chopper to hack the foreshaft into rough shape. At first he had problems trying to chop the ivory on the soft, cushioned surface of a rug. This problem was solved by using a large block of pressure-treated wood as an anvil. Long shavings curled off the ivory during the chopping process and were reminiscent of block plane shavings.

Once the general shape had been formed, he combined scraping, using the edges of the chopper and a blade struck from a blade core. A very porous chert abrader was used in rasp-like fashion to smooth out the chop-marks on the shaft.

The beveled end was finished by cutting gashes (basal roughening) into the face of the beveled platform and around the diameter of the shaft in the platform area. These gashes were made on the beveled end by using one edge of the elongate uniface chopper in knife-like fashion. The pointed end of the foreshaft was polished by using an obtuse flake ridge on the chopper for heavy rubbing followed by sand grit on leather, with the foreshaft stropped over the gritcovered surface.


Four foreshafts with lithic projectiles were constructed (see fig. 2). A Clovis (or Cloviswaisted) fluted point 6.98 cm long of Suwannee Limestone formation "cannonball" chert was hafted to an oak wood foreshaft with unwaxed dental floss and Elmers white glue. Total foreshaft-point weight 39.7 grams. This waisted Clovis fluted point (#l) was re-hafted to the ivory foreshaft mentioned below. A mastic of pine rosin and tallow was applied to the flute and to the beveled face of the foreshaft during hafting. Total foreshaft-point weight 53.1 grams.

A Simpson point 11.43 cm long of Ocala Group Gulf Coastal Plains chert from Bushnell, Florida was hafted to an alligator bone foreshaft 20.96 cm long using hafting mastic made from pine sap and animal tallow, and bound together by deer sinew. This was the heaviest of the foreshaft-point combinations at 90.3 grams. It withstood three or four throws before impacting the target at an awkward angle caused it to glance off and snap the lithic point at the intersection of the sinew binding. The sinew binding of the haft remained very secure around the basal fragment of the point that snapped. The sinew binding was later removed by letting the foreshaft soak in wan-n water for about an hour.

A parallel-sided "classic" Clovis fluted point 9.52 cm long of Suwannee Limestone formation Flint River chert was hafted to an ivory foreshaft 15.54 cm long using hafting mastic made from pine rosin and animal tallow, and bound by strands of bear-grass fibers. This foreshaft-point (#2) weighed 50.5 grams.


This initial experiment of tossing replica foreshafts tipped with Southeastern Paleoindian projectile points produced 53 throws which hit and stuck into the targets as summarized in the graphic “Summary of Results”


The purpose of this experiment was to duplicate the type of damage found on the assemblage of ivory artifacts from Florida proposed to be Paleoindian spear foreshafts. Even though there is no evidence that Paleoindians used the atl atl, it was used in this experiment because of the greater likelihood of producing impact damage over hand-thrown spears.

The foreshafts not only withstood 53 direct hits into rugged targets but also survived many more plunges into the dirt and, occasionally, into a hardwood tree. This suggests the foreshaft concept represents a functional design that works with little fear of damage. With most hits the point not only stabbed the palm tree, it also supported the weight of the spear shaft above the ground by its projectile point. On about a dozen occasions the lithic point had to be freed from the palm wood with a wood chisel and hammer.

Surprisingly, no damage was inflicted on the foreshaft replicas. Rather, the base of a Simpson point broke in its haft at the torque point where Webb and I had predicted the foreshaft would fail.

A problem of lesser importance was maintenance of the haft. The plant fiber haft quickly failed. The plant fiber hafting held up when the foreshaft penetrated beyond the haft, but the first time that the haft was left exposed, the weight of the spear shaft severed the plant fibers. The unwaxed dental floss stretched and wore and needed to be replaced or reinforced several times. The combination of animal sinew and mastic seemed to provide the most durable hafting.

The wood foreshaft tended to jam into the antler socket, especially on humid days. The ivory foreshaft jammed infrequently. On two occasions however it required pliers to dislodge the ivory shaft from the socket. Perhaps not coincidentally, these occurred on a humid day.

More testing will be conducted. We suspect that serious breakage occurred rarely, perhaps when spears hit bone.