Terminal Transferase (TdT) adds any deoxynucleotides in a template-independent manner to the 3' end of DNA.
People usually use this to poly-tail with a certain nucleotide. According to NEB's site, with an excess of a single nucleotide, the tail lengths with TdT after 15 minutes are dA: 300-500, dC: 100-150, dG: 15-25, dT: 250-350
However, here's some anecdotal evidence I have of nucleotide distribution: I used TdT for 30 minutes @37C in NEB4+CoCl2, with all 4 nucleotides (0.25mM each). Of 4 colonies sequenced, the random sequences were of lengths: 133, 198, 42, 69. The stats of the sequence generated was: 132 A, 54 T, 81 C, 175 G (total 442 nt). This shows a heavy G-bias which doesn't seem to agree with NEB. --Austin 19:35, 13 March 2006 (EST)