Macrovesicular steatosis (MaS), fibrosis, and inflammation have been associated with poor graft function after liver transplantation. We evaluated histological variation in livers to determine the optimal number of biopsies to estimate pathological characteristics in livers for transplantation. Specimens from autopsies performed during 3 months in 16- to 70-years-olds without known liver disease or drug and/or alcohol abuse were examined. Eight needle biopsies were performed, and hematoxylin and eosin–stained slides were evaluated. Percentages of MaS and microvesicular steatosis (MiS) were determined, and inflammation and fibrosis were scored as 0 to 4. MaS correlated positively with MiS, and inflammation correlated positively with fibrosis, whereas patient weight showed a significant correlation with liver weight and body mass index. No patient characteristic showed a significant correlation with histological findings. Subjects 55 years and older showed no increase in pathological findings compared with those younger than 55 years. When any site was compared with the average of the other sites, Spearman's ρ correlation ranged from 0.89 to 0.95 for MaS, 0.89 to 0.94 for MiS, 0.54 to 0.80 for inflammation, and 0.66 to 0.80 for fibrosis. Two biopsies explained 95% of variations for MaS and MiS and 85% for inflammation and fibrosis. There was no significant difference between findings in the right and left lobes of livers. These findings suggest that no single site best predicts pathological findings, and there is little variation among sites. In borderline cases of MaS, significant pathological characteristics may be found in additional biopsies. Therefore, we recommend two biopsy sites to evaluate donor livers with suspicious clinical histories.