The subgel phases of unlabeled, specifically chain perdeuterated and specifically 13C = O labeled representative samples of the n-saturated diacylphosphatidylcholines were studied by Fourier-transform infrared spectroscopy. Our results indicate that the spectroscopic properties exhibited by the subgel phases of the longer chain homologues are not consistent with that of a pure phase and we suggest that this is because the observed spectrum is a summation of spectroscopic features arising from both their subgel and L beta gel phases. Using spectral subtraction techniques, we obtained a spectrum which we believe is more representative of the pure subgel phase and from it we suggest that the subgel phase of the long chain phosphatidylcholines is an ordered crystallike structure containing two vibrationally inequivalent populations of lipid molecules arranged with the zigzag planes of their hydrocarbon chains parallel. For dipalmitoylphosphatidylcholine, our data indicate that its stable subgel phase is generally similar to that of the longer chain homologues but it is a more ordered structure in which the polar/apolar interfacial region is probably less hydrated. With the medium chain (N = 13-15) compounds, two populations of vibrationally equivalent molecules are also present in the subgel phase, but unlike DPPC and the longer chain homologues, the zigzag planes of their sn1- and sn2- acyl chains are perpendicular to each other, and a sn1-ester C = O group of one of the populations is in relatively close contact with an sn2-ester C = O group of the other population. With the shorter chain (N = 10 - 12) compounds, our data is indicative of a very complex quasi-crystalline assembly in which there may be at least three vibrationally inequivalent populations of lipid molecules with rotationally disordered hydrocarbon chains. Moreover, the conformation of the glycerol backbone may well be very different from that usually expected of this class of phospholipids. With all of these lipids, the structural pictures which emerge from our studies of the various subgel phases are in many aspects incompatible with that deduced from the single crystal x-ray studies of dimyristoylphosphatidylcholine. We suggest that this is because under our experimental conditions, these lipids have effectively been crystallized from water, whereas the sample used for the single-crystal x-ray study was crystallized from organic solvents.