The use of dual energy X-ray absorptiometry (DXA) is increasing in animal research, but data comparing whole body composition by DXA and chemical analysis (CHEM) in rats are limited. Lean and obese Zucker female rats were used to compare DXA (QDR1000W, Hologic, Waltham, MA) values with CHEM data for percent fat (%FATDXA, %FATCHEM), lean body mass (%PROTDXA, %PROTCHEM) and bone mineral content (%BMCDXA, %ASHCHEM). Four groups of rats (n = 9) were tested for differences in body composition due to consumption of a 100 g guar gum supplement/kg to see if DXA was as sensitive as CHEM in detecting body composition differences induced by diet. The study was analyzed using a split-plot ANOVA where the main plot was a 2 x 2 factorial with phenotype (obese or lean) and treatment (guar gum or control) as the effects, and the subplot was method of detecting body composition (DXA and CHEM), which was treated as a repeated measure. Absolute values for percent fat differed significantly (P < 0.0001) between the two methods as %FATDXA was consistently higher than %FATCHEM. There was not a statistically significant difference due to method for %PROT (P = 0.13). Values for %BMCDXA were significantly (P < 0.0049) lower than %ASHCHEM values. The differences in body composition due to diet treatment were detected similarly by DXA and CHEM. Significant correlations were found between the methods (P < 0.0001) for %FAT (r = 0.99), %PROT (r = 0.96) and %BMC or ASH (r = 0.81). Bland-Altman plots showed good agreement between methods, and regression equations were developed to estimate CHEM values from DXA readings. DXA may provide an alternative method for assessing changes in whole body composition.