Export citation and abstract BibTeX RIS
In the published version of this paper, the planet orbiting KIC 8435766 is referred to as "Kepler-XXb." The "XX" was intended to be a placeholder for an official Kepler number, which had not yet been assigned. While the article was in its final editorial stages, this planet was officially designated with the name "Kepler-78b" (see http://exoplanetarchive.ipac.caltech.edu/docs/KeplerNumbers.html). However, the authors neglected to communicate the new name to the editors before the article went to press. Thus, every time that "Kepler-XXb" appears in the text, in the figures, and in the tables, it should be read as "Kepler-78b."
Additionally, there was a minor error in the description of the manner in which the flux time series were normalized in Section 3.2. The recognition of this error does not alter the conclusions of the paper, but details are provided below to minimize confusion for other investigators who might analyze the same data in the future.
The text in Section 3.2 states that during the normalization of the flux time series, the data from each quarter were divided by the mean flux during that quarter. In fact, though, all of the data were divided by the maximum flux over the entire time series (including all of the quarters). Therefore, even after normalization, each quarter of data had a different mean flux, ranging from 0.92 to 1. The reason this is largely immaterial is that the data were subsequently analyzed with a model that included a "dilution parameter" specific to each Kepler season, i.e., by the 90° rotations of the spacecraft. This was a constant flux that was added to the theoretical model, which was then renormalized to unit during secondary eclipse, and then compared to the data. However, we note that the largest changes in flux with Kepler quarter are cyclic by Kepler season. By fitting for these seasonal dilution parameters, the mean fluxes of the various quarters were brought into agreement, though not by the exact method described in the paper.
The only (minor) difference between modeling the data as described, and as actually performed, is that the variations in mean flux across quarters were actually brought into agreement by adding small constant fluxes instead of by multiplying by factors close to unity. When the analysis was repeated as described in the published paper, the only affected parameters were those associated with the phase curve, with δocc = 10.1 ± 1.2 (0.33σ change) and with Aill = 4.2 ± 0.5 (0.4σ change).