On the period distribution of close-in extrasolar giant planets

Transit (TR) surveys for extrasolar planets have recently uncovered a population of "very hot Jupiters,'' planets with orbital periods of P <= 3 days. At first sight this may seem surprising, given that radial velocity ( RV) surveys have found a dearth of such planets, despite the fact that their sensitivity increases with decreasing P. We examine the confrontation between RV and TR survey results, paying particular attention to selection biases that favor short-period planets in TR surveys. We demonstrate that, when such biases and small-number statistics are properly taken into account, the period distributions of planets found by RV and TR surveys are consistent at better than the 1 sigma level. This consistency holds for a large range of reasonable assumptions. In other words, there are not enough planets detected to robustly conclude that the RV and TR short-period planet results are inconsistent. Assuming a logarithmic distribution of periods, we find that the relative frequency of very hot Jupiters (VHJs; P = 1 3 days) to hot Jupiters (HJs; P = 3 9 days) is similar to 10% - 20%. Given an absolute frequency of HJs of similar to 1%, this implies that approximately one star in similar to 500 - 1000 has a VHJ. We also note that VHJs and HJs appear to be distinct in terms of their upper mass limits. We discuss the implications of our results for planetary migration theories as well as present and future TR and RV surveys.