The chlorophyll a (Chl a) special-pair model of the primary donor of photosystem I (P700) does not account in a completely adequate fashion for the magnetic resonance properties observed for P700(+). Moreover, P700 is at least 420 mV easier to oxidize than is Chl a in vitro. Neither Chl a dimer formation nor selective ligation of Chl a can account for this potential difference. Enolization of the Chl a ring V beta-keto ester results in a very different pi electronic structure. The Chl a enol can be trapped as a silyl enol ether. In addition, the enol analog 9-desoxo-9,10-dehydro-Chl a can be prepared. Both the trapped enol and its 9-H analog are approximately 350 mV easier to oxidize than Chl a. The ESR spectrum of the cation radical consists of a single 6.1-G gaussian line that is line narrowed relative to that of Chl a(+) in a manner similar to P700(+). Electron-nuclear double resonance (ENDOR) spectroscopy resolves only a 3.5-MHz hyperfine splitting for the 3-methyl-group. The remaining splittings are all less than 3.5 MHz. The second moment of the ESR line of fully (13)C-enriched 9-desoxo-9,10-dehydro-Chl a(+) agrees with that of [(13)C]P700(+) to within 10%. Application of the special-pair model to the [(13)C]P700(+) second-moment data yields a 100% error. Ab initio molecular orbital calculations on ethyl chlorophyllide a enol cation bear out the ESR and ENDOR data. We conclude that a monomeric Chl a enol model provides a better description of the magnetic resonance parameters and oxidation potential of P700 than a Chl a special-pair model.