The GlobalSoilMap.net initiative requires the generation of continuous maps for soil properties, including pH. The standard method, ISO 10390, prescribes determining pH in a 1:5 suspension of soil in water, and CaCl2 pH in a 1:5 suspension of soil in 0.01 M CaCl2. The US Soil Survey employs a 1:1 suspension of soil in water and a 1:2 suspension of soil in CaCl2 for routine pH (Soil Survey Staff, 2009); more than a 100,000 soil samples have been analyzed in this format. Absent of a global standard method for determining routine pH and CaCl2 pH, the GlobalSoilMap.net project provides references for converting between methods (Aitken and Moody, 1991; Bruce et al., 1989; Conyers and Davey, 1988). These conversions were based on small, geographically narrowly defined sample sets of less than 100 samples. The extensive NSSC archive affords the opportunity to establish inter-method relationships representing any soil order, mineralogy, texture, or soil property, to reliably convert between methods, and to understand and define the influence of soil properties on any mutual bias between data sets that might be observed. A subset of soil samples from the National Cooperative Soil Characterization Database that were previously analyzed for 1:1 H20 and 1:2 CaCl2 (0.01 M) were selected for pH determination by 1:5 H2O and 1:5 CaCl2 (0.01 M). The samples represented all 12 soil orders 3 mineralogy classes, organic vs. mineral and 7 texture classes. For each category, between 25 and 30 samples were selected to represent a comprehensive pH range, requiring in the analysis of more than 750 samples. The regression analysis showed strong and significant relationships (R2>0.91) between methods, with slight differences due to soil orders, mineralogy, organic vs. mineral and grouped texture classes. The regression equation was pH 1:5 H2O = -0.51 + 1.06 pH 1:1 H2O with an RMSE = 0.44 pH units. Grouped texture classes had a significant influence on the relationship, but relatively small compared to the method of pH determination as shown by the F-Ratios of 8 and 3809, respectively. Adding the grouped texture classes to the prediction equation in a stepwise forward selection procedure resulted in a decrease of AIC from 676 to 647, indicating that the probability of loosing information by using only method of pH determination in the prediction is 5.0e-7 times. The simple model with method of pH determination variable only was selected for pH predictions.