We present a continuous time series on first cabin passenger fares for ocean travel from New York to the British Isles covering nearly a century of time. We discuss the conceptual and empirical difficulties of constructing such a time series, and examine the reasons for differences between the behavior of advertised fares and those based on passenger revenues. We find that while there are conceptual differences between these two measurements, as well as differences in the average values, the two generally moved in parallel, which means that the advertised fare series can serve as a reasonable proxy for movement of the revenue-based fares. We also find that advertised fares declined over time, roughly paralleling the drop in freight rates for U.S. bulk exports, until around 1890, but thereafter increased while freight rates continued to decline. We propose several hypotheses for this divergent behavior and suggest lines of future research.