Same as Fig. 9, but for the K I λ7699-Å line. We note that the excess absorption in the line core is probably seen at all eclipse phases maybe with the exception of the exact mid eclipse at around UT 5:07 (which is the noisiest spectrum of all though). The two tiny lines symmetric around K I, at λ7698.304 Å and 7699.486 Å, are two water-vapor lines that became dramatically stronger during umbral eclipse. The out-of-eclipse spectrum is now plotted with dots to additionally show the CCD pixel sampling of PEPSI in its R = 130 000 mode.