The　inhalation　is　one　of　important　exposure　ways　to　arsenic.　Traditionally,　the　health　risk　of　arsenic　exposure　from　particulate　matter　(PM)　was　assessed　by　using　total　arsenic,　which　may　erroneously　estimate　the　health　risk　of　arsenic　since　the　toxicity　of　arsenic　depends　on　its　chemical　species　and　not　all　arsenic　in　PM　is　bioaccessible.　Herein,　total　suspended　particles　(TSP)　were　collected　from　Taiyuan　in　China　during　whole　year　of　2018,　and　the　species　and　concentrations　of　arsenic　in　TSP　were　investigated　in　order　to　more　accurately　assess　the　health　risk　of　arsenic　exposure　from　TSP　and　evaluate　the　possible　sources　of　arsenic　in　TSP.　Total　arsenic　varied　within　1.16-28.4　ng/m(3)　with　a　mean　value　of　7.40　ng/m(3),　which　exceeded　the　standard　limit　of　China　(6　ng/m(3)).　Two　arsenic　species,　As5+　and　As3+,　were　detected　out　in　soluble　fractions　of　TSP,　with　As5+　as　dominant　species.　Total　arsenic,　soluble　arsenic　and　soluble　As5+　in　TSP　revealed　closed　correlation　each　other,　indicating　that　they　may　originate　from　similar　anthropogenic　and　crust　sources.　Soluble　As3+　showed　no　obvious　correlations　with　total　arsenic,　implying　that　soluble　As3+　has　different　dominant　sources.　The　ratio　of　As5+/As3+　significantly　varied　within　1.08-32.5　and　the　percentages　of　soluble　arsenic　in　total　arsenic　varied　within　50%-93%,　implying　that　arsenic　in　TSP　of　Taiyuan　has　multiple　sources　and　none　of　them　stably　dominated　during　2018.　Non-carcinogenic　risk　and　carcinogenic　risk　indicators　calculated　with　soluble　arsenic　species　showed　significant　difference　to　that　calculated　with　total　arsenic　or　soluble　arsenic　when　TSP　contained　equivalent　As5+　and　As3+,　verifying　that　it　is　necessary　and　more　accurate　to　assess　the　health　risk　of　arsenic　exposure　from　TSP　by　using　soluble　arsenic　species,　rather　than　total　arsenic　or　soluble　arsenic.　(C)　2021　Elsevier　B.V.　All　rights　reserved.