Taubenberger et al. have sequenced the polymerase genes of the pandemic 'Spanish' influenza A virus of 1918, thereby completing the decoding of the genome of this virus. The authors conclude from these sequences that the virus jumped from birds to humans shortly before the start of the pandemic and that it was not derived from earlier viruses by gene shuffling, a process called reassortment. However, we believe that their evidence does not convincingly support these conclusions and that some of their results even indicate that, on the contrary, the virus evolved in mammals before the pandemic began and that it was a reassortant. In light of this alternative interpretation, we suggest that the current intense surveillance of influenza viruses should be broadened to include mammalian sources.