8 Summary of observational clues

Before embarking on our discussion, we recap the highlights of our observations. First, the 2000 WET data shows eigenmodes from at least k=8 through k=19. We may also have detected the k=7 and k=20modes. However, their frequencies are similar to that of unrelated combination peaks, so their identification is less secure. For the first time, we have also found an $\ell=2$ mode in the GD 358 data; it is at $1255.4~\mu$Hz. Second, we see relatively few multiplet modes for a given k, with the exception of the k=8 and 9 modes. While the multiplet structure of the $\ell=1$ modes is muted, the combination peaks are enhanced to the point that we see combination modes up to 6th order. Combined with the previous WET runs, we see evidence for anticorrelation between the presence of multiplet structure and combination peaks. The presence of amplitude variability of the $\ell=1$ mode continues. In the August 1996 data, we saw the most extreme example yet, where all of the observed light was in a single (k=8 mode) for one night (which we call the "forte''). Data before and after the run show power in the nights before and after in other pulsation modes besides the k=8 and a much lower amplitude. The periods from the 1996 data are consistent with the 2000 data set, although there are differences in the details.