A magnetic reversal chronology has been obtained from a section over 700 m thick of radiometrically dated lava flows (7.9--11.5 m.y.) from the Hackberry Mountain area in central Arizona. There are an adequate number of polarity reversals in the Hackberry Mountain volcanic sequence to allow an apparent correlation with the sea floor record assigned to this time interval. A correction using a few radiometric ages widely spaced in the section for temporal control proves to be incorrect because it does not assess the length of hiatuses in the terrestrial sequence and assumes that the rate of volcanism and sedimentation is uniform. When a large number of stratigraphically controlled radiometric ages are used, temporal compression, extension, and shifting of parts of the Hackberry Mountain reversal pattern produce a different correlation with the ocean floor reversal pattern. In spite of such adjustments, not all of the stratigraphically controlled K-Ar ages and associated polarity reversals are in accord with ages estimated for polarity boundaries in the sea floor record. The Hackberry Mountain pattern can be less well fitted to the marine sediment polarity record than to the sea floor anamaly record, implying that there are unrecognized hiatuses in the marine sediment record. We conclude that a lack of apparent similarity in the adjusted 8- to 11-m.y.-old Hackberry Mountain pattern and the currently accepted reversal time scale derived from marine magnetic anomalies reflects (1) loss or absence of the geologic record in the terrestrial volcanic sequence, (2) perhaps brief reversals not recognized in the marine record, and (3) a possible lack of strict linearity in the assignment of estimated ages to the sea floor pattern, which is based on the assumption of uniform spreading rates. A reversal pattern obtained from a terrestrial volcanic sequence older than about 5 m.y. thus is of no value, in itself, for establishing a chronology.