The recently deciphered sequence of the human mitochondrial genome is analyzed in the light of an archigenetic hypothesis, according to which mitochondria are derived neither from pro- nor eukaryotes but from more primitive organisms. The possibility that animal mitochondria have only one gene both for elongator and initiator methionine tRNA is supported but C-A pair forming cytosine in the anticodon of these tRNAs is considered to be unmodified. The evolution of the gene and of the codon reading pattern of methionine tRNA is discussed.