Understanding the form in which gold is transported in surface- and groundwaters underpins our understanding of gold dispersion and (bio)geochemical cycling. Yet, to date, there are no direct techniques capable of identifying the oxidation state and complexation of gold in natural waters. We present a reversed phase ion-pairing HPLC-ICP-MS method for the separation and determination of aqueous gold(III)-chloro-hydroxyl, gold(III)-bromo-hydroxyl, gold(I)-thiosulfate, and gold(I)-cyanide complexes. Detection limits for the gold species range from 0.05 to 0.30 μg L-1. The [Au(CN) 2]- gold cyanide complex was detected in five of six waters from tailings and adjacent monitoring bores of working gold mines. Contrary to thermodynamic predictions, evidence was obtained for the existence of Au(III)-complexes in circumneutral, hypersaline waters of a natural lake overlying a gold deposit in Western Australia. This first direct evidence for the existence and stability of Au(III)-complexes in natural surface waters suggests that Au(III)-complexes may be important for the transport and biogeochemical cycling of gold in surface environments. Overall, these results show that near-μg L-1 enrichments of Au in environmental waters result from metastable ligands (e.g., CN-) as well as kinetically controlled redox processes leading to the stability of highly soluble Au(III)-complexes.