Evidence indicating the presence of axonal mRNA in mammalian neurons has been limited to cytoskeletal components such as beta-actin, tau, and beta-tublin, and is still a subject of controversial debate. A new strategy is needed to prove that axonal transport of mRNAs is a general phenomenon, and to determine the purpose of this transport. I have therefore established a method to comprehensively identify axonal mRNAs. Single axonal terminals of cultured cerebral cortical neurons were aspirated, and total RNA was prepared and served for RT-PCR using a tagged random primer. Nine kinds of mRNA (Axomer-1 to -9) were shown to be expressed in an axonal domain, none with any homology with any functionally known sequences. These results suggest that quite a few mRNAs transported in axons may exist in the central nervous system of mammals, or at least of rats. This method appears to be useful to identify mRNAs in axons inclusively, sequence-independently, and even at low expression levels.