Two KIWI-GA runs for R136a3. In blue the original run, having a setup identical to that of all other stars (resulting in T_eff = 42 kK). In red the run where we fixed T_eff to 50 kK, all other aspects of the setup held the same. For some lines the higher temperature results in a better fit than the original run (light red background), while for other lines the original run resulted in better fits (light blue background).