Oriented archaeomagnetic samples were collected from 158 in situ features at 33 sites in the southwestern U.S. Seventy-three independently dated features were used for analysis of secular variation. A moving-window smoothing technique with outlier rejection was developed to objectively compute a virtual geomagnetic pole (VGP) path. Weighted Fisher statistics, accounting for errors in both dating and archaeomagnetic direction, were used to compute a mean VGP and standard error (A95) within each window of time. The VGP moves from 86°N, 83°E at A.D. 750 to 74°N, 192°E at A.D. 1075, and then to 85°N, 236°E at A.D. 1425. Secular variation of the VGP ranges from .00423-.350°/yr with a median of .0359°/yr. The A95s for the VGP curve range from 1.33-5.28°. Archaeomagnetic declinations and inclinations in Tucson range from 346-359°E and 47-60°. The close agreement with a similarly constructed VGP curve for Arkansas implies a small nondipole field in North America between A.D. 1150-1450. The VGP curve is significantly different from that of R. L. DuBois. Three case histories of archaeomagnetic dating suggest the validity of the new curve. The Thellier-Thellier paleointensity experiment was performed on 187 specimens from 77 potsherds. A significant magnetic fabric anisotropy, typically 30%, was found. Using an easy-plane model of magnetization, the anisotropy correction systematically increased the paleointensities by 5%. Paleointensities were interpreted for 127 samples from 54 sherds. All ceramics were independently dated and used along with other results to construct a virtual axial dipole moment (VADM) curve. A moving-window smoother using weighted statistics within each interval of time and incorporating outlier analysis was used. The Southwestern VADM at 300 B.C. is 14 x 10('22)Am('2), decreases to 8.6 x 10²²Am² at A.D. 800, and increases to 12.9 x 10²²Am² at A.D. 1500. Paleointensities from Hohokam ceramics compare favorably with paleointensities from other Southwestern ceramics, regardless of whether the long Hohokam chronology beginning at 300 B.C. or the short chronology beginning at A.D. 500 are used. There is also reasonable agreement between the Southwest and Mesoamerican VADM records.

Paleomagnetic poles were obtained from the Chinle, Kayenta, Summerville, and Morrison Formations. Combined with paleomagnetic poles from the Moenave Formation, poles from the Chinle and Kayenta formations record ∼30 m.y. of North American apparent polar wander (APW) within a regional stratigraphic succession. During the Carnian and Norian stages of the Late Triassic, Chinle poles progress westward. During the Hettangian through Pliensbachian stages of the Early Jurassic, the pattern of APW changed to an eastward progression. Even after correction for 4° clockwise rotation of the Colorado Plateau, a sharp corner in the APW path (J1 cusp) is resolved near the pole from the Hettangian/Sinemurian (∼200 Ma) Moenave Formation (59.4°N; 59.2°E). The J1 cusp implies an abrupt change from counterclockwise rotation of Pangea prior to 200 Ma to clockwise rotation thereafter. Paleomagnetic poles obtained from the Summerville and Morrison formations are consistent with the Middle and Late Jurassic APW path described by the Corral Canyon and Glance Conglomerate paleomagnetic poles as well as a Late Jurassic Cusp (J2 cusp) in the APW path. The APW path described by the J2 cusp, a single Morrison Formation pole, and mid-Cretaceous paleomagnetic poles suggest from ∼150-126 Ma the North American plate experienced a minimum rate of motion of 0.93°/m.y. which is similar to rates calculated for the Late Triassic (0.73°/m.y.) and Jurassic (0.66°/m.y.). This rate is much lower than rates based on previous APW paths. Thermal demagnetization and data analysis indicate that within-site dispersion is an important criterion for selecting sites which retain a high unblocking-temperature, characteristic remanent magnetization (ChRM). This criterion was used to define at least three stratigraphically-distinct, antipodal polarity-zones within each member/formation, suggesting the ChRM was acquired soon after deposition. ChRMs from 15 to 22 sites in the Upper Shale Member of the Chinle Formation define an early Norian paleomagnetic pole position of 57.4°N, 87.8°E (K = 60, A₉₅ = 5.0°). ChRMs from 18 to 43 sites in the Owl Rock Member of the Chinle Formation define a middle Norian paleomagnetic pole position of 56.5°N, 66.4°E (K = 183, A₉₅ = 2.6°). ChRMs from 23 of 35 sites in the Kayenta Formation define a Pliensbachian pole position of 59.0°N, 66.6°E (K = 155, A₉₅ = 2.4°). ChRMs from 15 to 35 sites in the Summerville Formation define a late Callovian pole position of 53.8°N, 133.6°E (K = 25, A₉₅ = 7.5°). ChRMs from 15 sites in the Morrison Formation (9 from the study of Steiner and Helsley [1975]) define a single, ∼147 Ma, paleomagnetic pole position of 64.1°N, 152.4°E (K = 113, A₉₅ = 3.6°).