Bone Density and Survivorship
As explained in chapter 2, the archaeological assemblage that an analyst receives is a biased sample of the animals that were once present at the site. Some bones are so small that their recovery using standard archaeological techniques is unlikely. Others are so fragile that the field crew may not have been able to recover them without reducing them to small splinters of bone. Natural or cultural processes such as weathering or butchering may have destroyed some bone long before archaeological excavation began. Bones that were never collected by the field crew or never survived transportation from the field to the faunal analyst cannot be directly studied, but they are not completely out of reach of our analyses.
Bone density studies provide us with a means of measuring the density-mediated attrition of an assemblage. This is an estimation of the impact of bone density on the survivorship of the bones that make up an assemblage. Essentially, we can assess what has been lost by documenting what is present. Thanks to some important studies, recording bone density data is simple and can be done as the analyst catalogs each bone. It is not necessary to record this data for each and every bone; instead select one or more common species, such as deer, dog, or rabbit, and record the required data for each deer, dog, or rabbit bone as it is cataloged.
Survivorship is an analysis of those bones that did survive to make it into the faunal assemblage in order to evaluate what, if any, destruction occurred and resulted in the elimination of other bones from the assemblage. One way to study survivorship is through density-mediated attrition. Researchers have used medical equipment like that used to detect osteoporosis in humans to scan the bones of many animal species. Since a single bone does not have a uniform density, each bone is scanned multiple times using different orientations of the bone. The resulting density data is published in figure and table format (e.g., Lam et al. 1999, 2003; Novecosky and Popkin 2005; Pavao and Stahl 1999; Stahl 1999) showing what portion of the bone was scanned along with the density reading for that scan site (fig. 1). 
Figure 1. Examples of scan sites used to analyze density-mediated attrition. (Adapted from Novecosky and Popkin 2005, Pavao and Stahl 1999, and Stahl 1999).
To include bone density data in your faunal catalog, simply determine if the scan sites for a particular bone are present when you identify that bone. For example, if you are recording density data for rabbit bones and identify a rabbit femur, you would consult a diagram showing the different scan sites of rabbit femurs. Once that image has been obtained (you should have these images on hand when you start cataloging), compare it to the archaeological femur. If the proximal femur is missing, then you don’t have those scan sites; record a value of 0 for those. If a mid-shaft scan site is only half present, record a value of 50 for it. If all distal articulation scan sites are present, record values of 100 for each. These are your survivorship values for each scan site.
When you are ready to analyze the bone density data, you need to add up the survivorship values for each scan site separately. Some scan sites will occur only once in a skeleton, such as a scan site of the sacrum, while others will occur more than once, such as a scan site on the proximal femur. So to normalize the data, divide the total percent present for each scan site by the number of times that scan site occurs in a single skeleton of that species. Then divide the resulting value by the same data for the most commonly represented scan site for that species. The result is a normalized percent survivorship statistic (Lyman 1994: 256). Then plot the percent survivorship against the bone mineral density value for each scan site that you cataloged. The resulting graph might resemble figure 2 or 3.
Calculating Percent Survivorship for Bone Density Studies
Percent Survivorship = (total percent present of each scan site ÷ number of times it occurs in a skeleton) ÷ (total percent of most common scan site ÷ number of times it occurs in a skeleton)
Figure 2. Percent survivorship of white-tailed deer (Odocoileus virginianus) from a prehistoric site in New York. This graph shows the presence of density-mediated attrition as the percent survivorship increases as bone density increases.
Figure 3. Percent survivorship of domestic dog (Canis familiaris or Canis lupus familiaris) from a prehistoric site in New York. This graph shows no clear pattern of density-mediated attrition, as the percent survivorship does not increase as bone density increases. Many of the dogs from this site were formally buried in their own graves.
Conduct this analysis for each species for which density data is available. This will allow you to determine if certain animals were treated differently in death than others, by being more or less exposed to density-mediated attrition. Animals that are intentionally buried soon after death are less likely to have suffered density-mediated attrition than those that were discarded into a surface midden.