In order for in vitro methods to become widely accepted as tools that accurately assess soil arsenic (As) exposure through the oral ingestion pathway, a better understanding is needed regarding which fractions of soil As are being measured in the in vitro extraction. The objective of the current study is to (1) identify in vitro bioaccessible (IVBA) and non-IVBA fractions of soil As using sequential extraction; and (2) determine the sorptive phases of soil in non-IVBA As soil fractions. Nineteen soils with a range of soil properties were spiked with 250 mg/kg of sodium arsenate and aged. In vitro bioaccessible As (IVBA As) was then determined using The Ohio State University in vitro gastrointestinal method (OSU-IVG), and soil As was fractionated using sequential extraction into: (F1) non-specifically sorbed; (F2) specifically sorbed; (F3) amorphous and poorly crystalline oxides of Fe and Al; (F4) well-crystallized oxides of Fe and Al and residual As phases. The IVBA As across the 19 soil ranged from 0.36 to 2.75 mmol/kg (12 to 86%) with a mean of 1.26 mmol/kg (42%) in the gastric phase and from 0.39 to 2.80 mmol/kg (13 to 87%) in the intestinal phase with a mean of 1.32 mmol/kg (43%). The results of the sequential extraction showed that IVBA As extracted by the OSU-IVG is the As present in the first two fraction (F1 and F2) of the sequential extraction. In the non-IVBA fractions, highly significant relationships (P < 0.01) exist between F3 As and log transformed F3 Fe (r (2) = 0.74), but not F3 Al. In addition, the gastric extraction dissolves a significant fraction of soil Al, but not soil Fe, therefore As sorbed to Al oxides likely contributed to IVBA As and is accounted for in the F2 fraction of the sequential extraction. In vitro methods that demonstrate the ability to extract the similar soil fractions that occur in vivo across a wide range of soil types and As-contaminant sources is an important criteria for in vitro method validation. Further research that includes soils with multiple As-contaminant sources (mining, pesticide, etc.), soil As fractionation, and in vivo bioavailability is needed in order to determine if F1+F2 are the bioavailable As fractions in soils that vary in total As content and sorbed As species.