In that year, a couple of guys named Nirenberg and Matthaei, working at the NIH, used our taxpayer dollars and some synthetic RNA to pry the puzzle apart. They made an mRNA molecule that contained only uracil, which they called “poly-U.”  They fed poly-U to an in vitro translation mixture, and analyzed the resulting protein. What they got was poly-phenylalanine—a protein made of only one amino acid. So it appeared that U coded for phenylalanine. Of course, with only four bases, the correspondence between amino acids and nucleotides couldn’t be one to one. So was it UU? Nope—if there’s four bases used to form a code of doublets, you can only make 4² = 16 words. Not enough to code for 20 amino acids. So was it  UUU? That would be enough. In fact, it was more than enough. So redundancy wasn’t out of the question. Could it then be UUUU? Or UUUUU? I trust U get the picture. Still, this was exciting. Nirenberg and Matthaei cooked up poly-C  and poly-A next, and got peptides made up of just proline and lysine, respectively. Poly-G proved a little more difficult, but by now a pattern was emerging. A bit more fiddling around demonstrated that it did in fact take only three nucleotides to code for an amino acid. No more, no less. Three must be the number of the nucleotides, and the number of the nucleotides must be three. Four nucleotides does not work, neither worketh two, unless thou then proceed to three.