Excitation energy transfer in DNA has similarities to charge transfer, but the transport is of an excited state, not of mass or charge. Use of the fluorescent, modified adenine base 2-aminopurine (2AP) as an energy trap in short (3- to 20-base) single- and double-stranded DNA oligomers is reviewed. Variation of 2AP's neighboring sequence shows (1) relatively efficient transfer from adenine compared to that from cytosine and thymine, (2) efficient transfer from guanine, but only when 2AP is at the 3′ end, (3) approximate equality of efficiencies for 3′ to 5′ and 5′ to 3′ directional transfer in adenine tracks. The overall, average transfer distance at room temperature is about four adenine bases or less before de-excitation. The transfer fluorescence excitation spectral shape is similar to that of the absorption spectrum of the neighboring normal bases, confirming that initial excitation of the normal bases, followed by emission from 2AP (i.e. energy transfer), is occurring. Transfer apparently may take place both along one strand and cross-strand, depending on the oligomer sequence. Efficiency increases when the temperature is decreased, rising above 50% (overall efficiency) in decamers of adenine below -60°C (frozen media). Modeling of the efficiencies of transfer from the nearest several adenine neighbors of 2AP in these oligomers suggests that the nearest two neighbors transfer with near 100% efficiency. As bases in B DNA, as well as in single-stranded DNA, are separated by less than 5 Å (less than the size of a base), standard Förster transfer theory should not apply. Indeed, while both theory and experiment show efficiency decreasing with donor-acceptor distance, the experimental dependence clearly disagrees with Förster 1/r6 dependence. It is not yet clear what the best theoretical approach is, but any calculation must deal accurately with the excited states of bases, including strong base-base interactions and structural fluctuations, and should reflect the increase of efficiency with temperature decrease and the relative insensitivity to strandedness (single, double). Attempts to use DNA as a molecular "fiber optic" face three primary challenges. First, reasonable efficiency over more than a base or two occurs only in adenine stretches at temperatures well below freezing. Second, transfer in these adenine tracks is efficient in both directions. Third, absorption of UV light occurs randomly, making excitation at a specific site on this "fiber optic" a challenge. © 2007 American Society for Photobiology.