Triangular hafted bifaces of different sizes constituted the majority (n=94, 53%) of the artifacts in the sample. They were separated into large (max. length >3.3 cm) and small (max. length <3.3 cm) categories, as size can potentially distinguish temporal or functional traits. The majority (n=78, 83%) of the triangular types were in the smaller size category with lengths between 1.8 and 3.3 cm. Virtually all of the smaller points were probably true arrowheads. Many of the larger specimens might represent hafted knife blades or projectiles. Larger triangular blades similar to the ones found at Jamestown have been recovered elsewhere in Virginia, including the Shannon Site in Montgomery County. Found there in association with slotted antler handles (Benthall 1969), indications are that they served often as cutting tools rather than weapons.

Several lithic raw materials were represented among the triangular hafted bifaces (see Appendix A). The 16 large triangular specimens were strongly dominated by quartzite (81%) with only two examples made of quartz and one of dark chert. A more diverse range of raw materials was used for manufacture of the smaller arrow points. Locally available quartzite again dominated the sample (51%), but quartz (26%), jasper (14%), dark chert (5%), orthoquartzite (3%), and metavolcanic materials (1%) were represented. Like quartzite, quartz commonly occurs locally in pebble or cobble form, and occasionally metavolcanic cobbles of argillite, rhyolite, or related material can be found in local stream gravels. Jasper, dark chert, and orthoquartzite, on the other hand, are extremely rare if not altogether absent in local gravels. Their sources are known to occur some distance away, specifically in the outer Coastal Plain for pebble jasper (Hodges 1998:60, Geier 1990), in the mountainous Appalachians for dark chert, and northeastern North Carolina for orthoquartzite (Bottoms 1968). The occurrence of jasper, dark chert, and orthoquartzite projectile points at other reported late prehistoric/ protohistoric sites along the James River from Jamestown to Richmond is very low (Hodges and Hodges, Ed. 1994; McLearen and Mouer 1994).

Patterns of failure, or breakage, among these artifacts can be telling with respect to function. For the purpose of the analysis presented here, the location and frequency of breaks were recorded for each artifact (see Appendix A). Among the large triangular hafted bifaces, 19% (n=3) were unbroken. The broken examples all exhibited failure in only one location rather than in multiple places. The most common damage was a lateral snap across the midsection (44%) which could result from the stress of prying and heavy cutting as readily as it could from an impact. Generally speaking, the lack of massive failure supported the notion that these larger triangular blades served as knives as often as projectiles.

Of the smaller arrow points, 18% were not broken. Among the remainder, failure of exactly half the points was relatively radical as represented by multiple breaks. Failure most commonly occurred in these cases (63%) at the distal tip or midsection and at one corner. Such massive collapse was experimentally demonstrated to be typical of points used as projectiles (Flenniken and Raymond 1986). Distal damage, meaning breakage at the tip or across the midsection, was the single most common type of damage among the small points (55%), which is also a pattern characteristic of used arrow points.

When failure patterns were examined relative to raw material type, notable contrasts emerged. Specifically, the small triangular points made of jasper and dark chert, both non-local materials, have been recovered intact significantly more often (47%) than those made of other materials (11%). Of the eight points made of these materials that are damaged, only one exhibited damage at the tip and only one had failed in more than one place. In both instances the damage was very slight. The relative lack of damage to these points made of higher quality, non-local materials might indicate special handling or a unique origin in the early 17th-century fort context.

Linking lithic type with failure frequency raises the issue of differential tensile strengths. Based on Callahan's scale of lithic toughness, basic chert and jasper are more prone to breakage than quartzite (Callahan 1979:16). Since they are more brittle and less durable, chert and jasper failure rates are expected to be higher than quartzite. The Jamestown Rediscovery assemblage sample includes evidence to the contrary, suggesting that some other phenomenon was responsible for the archaeological correspondence between lithic type and breakage.