The ambient tax is often considered as an e¢ cient instrument to achieve a …rst best outcome of ambient pollution when the regulator is less informed than the polluters. Since the ambient tax was never implemented in the …eld, empirical evidence is missing. Available experimental …ndings provide mixed evidence: e¢ ciency is higher under external damage, i.e. if ambient pollution a¤ects non-polluters (Spraggon, 2002(Spraggon, , 2003 than under internal damage, i.e. if ambient pollution a¤ects polluters themselves (Cochard et al., 2005). Since these two types of experiments relied on very di¤erent designs, it is worthwhile to compare them under a common experimental design. Our main …nding is that the ambient tax is equally e¢ cient under external damage than under internal damage.