Making use of a combination of ab initio calculated geometries, orbital energies, and orbital spatial distributions as well as experimental information about bond lengths, bond energies, vibrational frequencies, and dipole moments, the nature of the terminal PO bond in phosphates such as (MeO)(3)PO was probed and compared to the case in MeO-P = O where P is trivalent and a PO pi bond is thus assumed to exist. We find that the MeO-P and terminal PO bond lengths in (MeO)(3)PO are essentially the same as in MeO-P = O and the terminal PO lengths are substantially shorter than single P-OMe bond lengths. We also find that the HOMO orbital energies in the two compounds are within 0.1 eV of one another and that these orbitals have spatial characteristics much like one would expect of a bonding pi orbital connecting two atoms from different rows of the periodic table. Using this data, making a comparison to the more familiar bonding arising in N-2, CO, and BF, and taking note of the dipole moments in compounds known to possess dative bonds, we conclude that it is best to represent the terminal PO bond in phosphates in terms of valence-bond structures such as (MeO)(3)P = O in which the formal charges are (PO0)-O-0 and where a single PO pi bond exists. However, when it comes to characterizing the PO antibonding pi* orbitals, significant differences arise. Electronic structure methods were able to identify the pi* orbital of MeO-P = O and to determine its energy (the MeO-P = O- anion is even bound). Similar attempts to identify the PO pi* orbital in the unbound (MeO)(3)P=O- anion lead us to conclude that this anion state is probably so strongly coupled to the continuum (i.e., to states corresponding to (MeO)3P=O plus a free electron) that it is so short lived as to be undetectable in experiments.