(Supplementary Table 1) Mercury isotope values for common and thick-billed murre (Uria aalge, Uria lomvia) eggs from Alaska

Atmospheric deposition of mercury to remote areas has increased threefold since pre-industrial times. Mercury deposition is particularly pronounced in the Arctic. Following deposition to surface oceans and sea ice, mercury can be converted into methylmercury, a biologically accessible form of the toxin, which biomagnifies along the marine food chain. Mass-independent fractionation of mercury isotopes accompanies the photochemical breakdown of methylmercury to less bioavailable forms in surface waters. Here we examine the isotopic composition of mercury in seabird eggs collected from colonies in the North Pacific Ocean, the Bering Sea and the western Arctic Ocean, to determine geographical variations in methylmercury breakdown at northern latitudes. We find evidence for mass-independent fractionation of mercury isotopes. The degree of mass-independent fractionation declines with latitude. Foraging behaviour and geographic variations in mercury sources and solar radiation fluxes were unable to explain the latitudinal gradient. However, mass-independent fractionation was negatively correlated with sea-ice cover. We conclude that sea-ice cover impedes the photochemical breakdown of methylmercury in surface waters, and suggest that further loss of Arctic sea ice this century will accelerate sunlight-induced breakdown of methylmercury in northern surface waters.

(Table 2) Analyses of prominent ash layers recovered at DSDP Site 18-178, Gulf of Alaska

A sequence of ash layers recovered from site 178 of the Deep Sea Drilling Project in the Gulf of Alaska was studied to determine the nature of highly explosive volcanic eruptions associated with the Aleutian Arc and Alaskan Peninsula during the last 8 m.y. The major-element chemistry of 25 distinct ash layers was determined. When the analyses are plotted on conventional major-element variation diagrams, the unusual, highly evolved, calc-alkalic characteristics of the ashes are apparent. Perhaps more significantly, there is a good correlation of certain indices of the degree of chemical evolution of each ash (SiO2 content and Larsen index) with sample age. Both parameters vary cyclically, with maximum values of both indices occurring at present, 2.5, and about 5.0 m.y. ago. The cause of the cyclic activity, as well as discontinuous volcanic activity reported for other areas by other investigators, is not precisely known. However, we suggest that variable rates of subduction provide a viable hypothesis for discontinuous volcanic activity associated with convergent plate boundaries.