PURPOSE: To establish the reliability of anthropometric and dual energy X-ray absorptiometry(DXA) techniques used to assess percent body fat (% BF)in female athletes; to establish limits for detecting the smallest real change in % BF associated with anthropometric and DXA testing;to evaluate the validity of commonly used % BF prediction equations recommended by national certification programs along with equations derived from Multicompartment (MC), and DXA, in female athletes; and to create a new DXA based regression equation for elite female athletes. METHODS:Female athletes aged 17-31 were recruited into the study and participated in the establishment of anthropometric reliability (N=20), DXA reliability (N=32), and /or skinfold validity (N=95) testing. Anthropometric testing consisted of measurements of skinfolds,circumferences, and breadths. DXA measurements were conducted using a GE Lunar Prodigy DXA which served as the criterion measure (% BF DXA). RESULTS: Excellent reliability for both anthropometric sum5 skinfolds (ICC= .997, %TEM=0.9 %) and DXA (ICC =.996, CV =1.13% BF) techniques allows for detection of smallest real differences of 2.2 mm and 721g in summed skinfolds (sum5) and fat mass respectively. The DXA based equation of Ball et al.(2004) displayed the greatest validity of existing equations R=.874, total error (TE) 2.9% BF, and Bland Altman Limits of Agreement -4.7to 6.5 % BF. The newly created regression equation demonstrated a non-linear characteristic and displayed similar predictive ability R= .840, TE 3.0%BF, and Bland Altman Limits of Agreement of -6.1to 6.1 % BF. CONCLUSIONS: Anthropometric equations derived from various criteria yielded dissimilar results. Long utilized popular equations advocated in national accreditation schemes (ACSM, CSEP) show considerable bias compared to modern values obtained by current DXA technology. A new regression equation was created for female Canadian athletes 17-31yrs of age using skinfolds taught in the Canadian national professional certification program (CSEP).