Abstract：The wide-spaced sampling method is widely used to delineate large-scale geochemical patterns. However, the premise is that the large-scale anomalies must be reproducible in the follow-up surveys. A study area of 6400 km2 in Xinjiang was surveyed to investigate the reproducibility of geochemical anomalies delineated by different sampling densities, i.e. 1 sample/ 100 km2, 1 sample/25 km2 and 1 sample/4 km2. The results show that (1) the background values and average values obtained by different sampling densities are very consistent, (2) the large-scale geochemical patterns are very similar and reproducible by different sampling densities,（3）the values are more scattered as the sampling spacing becomes smaller, and (4) ultra-low (1 sample/ 100 km2) and very low-density (1 sample/25 km2) sampling can only delineate large-scale geochemical anomalies, but low-density sampling (1 sample/4 km2) can delineate both large-scale anomalies and small ones.