Perchlorate (ClO(4)(-)) has been detected widely in groundwater and soils of the southwestern United States. Much of this ClO(4)(-) appears to be natural, and it may have accumulated largely through wet and dry atmospheric deposition. This study evaluates the isotopic composition of natural ClO(4)(-) indigenous to the southwestern U.S. Stable isotope ratios were measured in ClO(4)(-) (delta(18)O, Delta(17)O, delta(37)Cl) and associated NO(3)(-) (delta(18)O, Delta(17)O, delta(15)N) in groundwater from the southern High Plains (SHP) of Texas and New Mexico and the Middle Rio Grande Basin (MRGB) in New Mexico, from unsaturated subsoil in the SHP, and from NO(3)(-)-rich surface caliche deposits near Death Valley, California. The data indicate natural ClO(4)(-) in the southwestern U.S. has a wide range of isotopic compositions that are distinct from those reported previously for natural ClO(4)(-) from the Atacama Desert of Chile as well as all known synthetic ClO(4)(-). ClO(4)(-) in Death Valley caliche has a range of high Delta(17)O values (+8.6 to +18.4 per thousand), overlapping and extending the Atacama range, indicating at least partial atmospheric formation via reaction with ozone (O(3)). However, the Death Valley delta(37)Cl values (-3.1 to -0.8 per thousand) and delta(18)O values (+2.9 to +26.1 per thousand) are higher than those of Atacama ClO(4)(-). In contrast, ClO(4)(-) from western Texas and New Mexico has much lower Delta(17)O (+0.3 to +1.3 per thousand), with relatively high delta(37)Cl (+3.4 to +5.1 per thousand) and delta(18)O (+0.5 to +4.8 per thousand), indicating either that this material was not primarily generated with O(3) as a reactant or that the ClO(4)(-) was affected by postdepositional O isotope exchange. High Delta(17)O values in ClO(4)(-) (Atacama and Death Valley) are associated with high Delta(17)O values in NO(3)(-), indicating that both compounds preserve characteristics of O(3)-related atmospheric production in hyper-arid settings, whereas both compounds have low Delta(17)O values in less arid settings. Although Delta(17)O variations in terrestrial NO(3)(-) can be attributed to mixing of atmospheric (high Delta(17)O) and biogenic (low Delta(17)O) NO(3)(-), variations in Delta(17)O of terrestrial ClO(4)(-) are not readily explained in the same way. This study provides important new constraints for identifying natural sources of ClO(4)(-) in different environments by multicomponent isotopic characteristics, while presenting the possibilities of divergent ClO(4)(-) formation mechanisms and(or) ClO(4)(-) isotopic exchange in biologically active environments.