Albanese The Study of Secular Change Using Skeletal Data 139 A Critical Review of the Methodology for the Study of Secular Change Using Skeletal Data John Albanese Secular changes or secular trends are non-genetic changes that occur over multiple generations in a population. The changes are not due to evolution because there are no corresponding changes in allele frequencies in the population. These changes typically correspond with changes in living condi- tions and are reflected in the population means for growth and development, and not necessarily in any one individual. The most studied secular change is change in stature or height. Improvements in living conditions have resulted in measurable increases or positive secular changes in stature from parents to children. The most obvious cases have been observed when comparing first genera- tion immigrants to their children. Conversely, decreases or negative secular changes in mean population stature from parents to children have been documented when living conditions worsen due to natural disasters, wars, or prolonged eco- nomic hardships. Several sources of data can be used for the study of secular change in stature and body size includ- ing historical stature data (Steckel 1994) such as recruitment records (for example, Floud 1994; Mokyr and Gráda 1994) and stature measure- ments collected on cross-sectional samples specifi- cally for research on stature (for example, Brauer 1982; Tobias 1986). Skeletal data are underuti- lized yet potentially useful sources of information for the study of secular change in stature as well as body proportions. With a few exceptions (Angel 1976; Jantz 2001; Jantz and Meadows Jantz 2000; Meadows Jantz and Jantz 1999; Ousley and Jantz 1998; Tobias 1986; Tobias and Netscher 1977; Trotter and Gleser 1951) osteometric data have not been used extensively for this purpose. In all of these studies a similar methodology has been used to investigate secular change using skeletal data. The purpose of this paper is to critically review this methodology and underlying theoreti- cal concepts using empirical data (femur length) from three reference collections: the Robert J. Terry Anatomical Collection, the Coimbra Identified Skeletal Collection and the Forensic Anthropology Data Bank. The primary focus will be to investigate the effects of the current standard approach of combining samples from different reference sources into racial groups for the study of secular change. Because of the similarities among the sources of data, close comparisons are made between the results from the current study and Meadows Jantz and Jantz (1999), and to a lesser extent with Trotter and Gleser (1951) and Angel (1976). Investigation of Secular Change Using Osteometric Data Long bone length data can be used to assess stat- ure and secular change. Using in vivo stature measurements and long bone measurements, Trotter and Gleser (1951) demonstrate that trends derived from stature data and long bone data are synchronized. However, the more robust approach is to use long bone length data since the derived stature will result in additional sources of error inherent in stature estimation equations and cor- rection for the effects of age (Tobias 1986; Trotter and Gleser 1951). Long bone data may be supe- rior to historical stature data for the following reasons: 1) Changes in limb and body proportions can be investigated along with absolute changes in any given skeletal element (Angel 1976; Meadows Jantz and Jantz 1999). Although stature measurements have been collected for centuries as a biometric component of per-