The first spectroscopic campaign on a PG 1716 variable (long-period pulsating subdwarf B star) has successfully detected low-level velocity variations due to g-mode pulsations in PG 1627+017, superimposed on the known orbital motion. The strongest velocity mode is barely detectable at 1.0-1.5 km s-1, although the target is one of the brightest (V = 12.9) and largest amplitude (∼0.03 mag) stars in its class. Forty nights of observations on 2 m class telescopes in Arizona, South Africa, and Australia provided 84 hr of time-series spectroscopy over a time baseline of 53 days, with typical velocity errors of 5-6 km s-1. The derived radial velocity amplitude spectrum, after subtracting the orbital component, shows three potential pulsational modes 3-4 σ above the mean noise level of 0.365 km s-1, at 7201.0 s (138.87 μHz), 7014.6 s (142.56 μHz), and 7037.3 s (142.10 μHz). Only one feature is statistically likely to be real, but all three are tantalizingly close to, or a 1 day alias of, the three strongest periodicities found in the concurrent photometric campaign. The velocity spectrum also shows an unexpected component at twice the orbital frequency of PG 1627+017, possibly evidence of a slightly elliptical orbit, supporting Edelmann et al.'s recent results for other short-period sdB binaries. We further attempted to detect pulsational variations in the Balmer line amplitudes. The single detected periodicity of 7209 s, although weak, is consistent with theoretical expectations as a function of wavelength, and it rules out a degree index of l = 3 or 5 for that mode. Given the extreme weakness of g-mode pulsations in PG1716 stars, we conclude that future efforts will require larger telescopes, higher efficiency spectral monitoring over longer time baselines, improved longitude coverage, and increased radial velocity precision. © 2006. The American Astronomical Society. All rights reserved.
