Geocentric Dipole Field Research Articles

The Iceland Basin excursion: Age, duration, and excursion field geometry

AbstractThe Iceland Basin geomagnetic excursion coincided with the marine isotope stage (MIS) 6/7 boundary. The age and duration of the excursion, at seven North Atlantic sites with sufficient isotope data, are estimated by matching marine isotope stage (MIS) 7a–7c to a calibrated template. Two criteria for defining the excursion, virtual geomagnetic pole (VGP) latitudes <0° and <40°N, yield excursion durations of 1–4 and 2–5 kyr, respectively. The midpoints of the excursion are in the 189–192 ka range, with a mean of ∼190.2 ka. Although component magnetization directions are generally well defined, rapid changes in field direction during a time of low field intensity are not adequately recorded. During the excursion, VGPs transit southward over Africa and the South Atlantic, reach high southern latitudes at the culmination of the excursion, with partial recovery in relative paleointensity (RPI), and then track northward through the western Pacific. The high southern latitude VGPs, and the recovery in RPI, imply that the Earth's main axial dipole reversed polarity during the excursion, if only for ∼1 kyr; implying that excursions can be manifested globally and are important in millennial‐scale stratigraphy. VGP clustering in the South Atlantic and NW Pacific roughly coincide with maxima in the vertical‐downward component of the modern nondipole (ND) field determined at the Earth's surface, which implies that the ND field became dominant as the geocentric dipole field weakened during the excursion, and also that the ND field configuration is long‐lived on multimillennial timescales.

The Mono Lake excursion recorded in phonolitic lavas from Tenerife (Canary Islands): Paleomagnetic analyses and coupled K/Ar and Ar/Ar dating

We present a coupled paleomagnetic/dating investigation conducted on three different lava flows from the island of Tenerife (Canary Islands; Spain) erupted during the Mono Lake excursion (MLE). Paleomagnetic analyses consist in zero field demagnetizations (AF and/or thermal) and of Thellier and Thellier experiments using the PICRIT-03 set of criteria to select reliable intensity determinations. One of the flows is characterized by a direction largely deviated from the one expected from an axial geocentric dipole (GAD) field. Its paleointensity value is very low (7.8 μT). The two other sites are characterized by inclinations slightly shallower than the GAD value and by low intensity values (about 12 and 21 μT; present value: 38 μT). The three K/Ar ages combined with two 40Ar/ 39Ar ages range from 32.0 to 33.2 ka and they are not statistically distinguishable from one another. It therefore appears that these lavas have recorded the MLE (the only excursion in this time interval) confirming its brief duration (shorter than the minimum age uncertainties available). The mean age is younger but, within the uncertainties, consistent with the age of the 10Be peak and of the marine intensity low when reported in the most recent ice age model. These new results are the first ones with radiometric dating produced from the northern hemisphere. Combined with existing cosmogenic, marine and volcanic paleomagnetic data, these results are discussed in terms of dating, and geometry of the earth magnetic field during the excursion.

On Detecting Changes in the Earth's Radius

Summary Estimates of the Earth’s radius in the geological past can be made from paleomagnetic evidence. A method appropriate to the spherical environment of the data for dealing with this problem is given, which is applied to Devonian, Permian and Triassic data from Europe and Siberia, yielding estimated radii for these periods of 1.12, 0.94 and 0.99 times the present radius respectively. These estimates are not considered to be significantly different from the present radius. Egyed (1960) has suggested that paleomagnetic data could be used to find probable ancient radii for the Earth; the idea is that the mean direction of magnetization of rocks of the same geological age from widely separated collecting sites should be consistent with a geocentric dipole field. The problem is of interest since it has been suggested on geological and cosmological grounds that the Earth’s radius has changed greatly in the geological past. For example, Egyed (1956, 1957) requires a change of z per cent since the Paleozoic, and Carey (1958) a change of 45 per cent in the same time. Cox and Doell (1961) have applied Egyed’s method of calculation to Permian data from Europe and Siberia and gave an estimated Permian radius of 0.991 times the present radius. In this article a generalization of Egyed’s method of calculation is given, which is applied to Devonian and Triassic data as well as Permian data from this region into which some recent values not available to Cox and Doell have been incorporated, yielding estimates of the ancient radii for these three geological periods not significantly different from the present radius. Triassic and Carboniferous data from the North American continent have also been studied, but as yet there are not enough data from this continent to yield a meaningful result. The method given here, like that of Egyed, is based on the assumptions that during any change in the Earth’s radius, the size of each continent remains constant, and that during the particular geological age being studied, the position of the pole has not changed markedly, together with the usual assumptions involved in paleomagnetic studies; that the Earth’s field at the time in question may be represented by a geocentric dipole, and that it is possible by a study of rocks from a given rock unit to determine the approximate direction of the Earth’s field at that place at a known time. The set of relevant observations from a rock unit consists of four quantities, which are average values for the rock unit as a whole: the mean longitude +$ and latitude A$ of the rock unit, and the declination Di and inclination It of the estimated

The Geomagnetic Field in Permo-Triassic Time

Summary A least-squares fit of the 500-fathom or 1000-metre contour of the principal continental masses has been constructed which may approximate to the arrangement of the continents in Permo-Triassic time. Reliable PermoTriassic palaeomagnetic data have been plotted on this reassembly, mean pole positions for the larger fragments have been calculated, and grand mean poles for Permian and Triassic time have been estimated. To a first approximation results are consistent with a geocentric dipole field, but small systematic departures have been detected. These may be attributed to superimposed multipole components coaxial with a dipole field, although other explanations, such as anomalies originating in the crust cannot be completely excluded. Tests for the existence of multipole components are suggested. The effect of such components on the accuracy of continental reassemblies based on palaeomagnetism is briefly noted.

Magnetic Position Determination by Homing Pigeons?

How pigeons return home from unfamiliar release sites is a long-standing puzzle in animal behaviour. Walker (1998, 1999) has described a “vector summation model” which “identifies a novel coordinate that pigeons could use with magnetic total intensity to determine position”. The model is not applicable in a magnetic field generated simply by a geocentric dipole, but requires a field perturbed by higher-order sources. Tests are devised to simulate the addition of both regional and local magnetic anomalies to a geocentric dipole field, and to calculate the directions of the home loft from a number of release sites. The results indicate that a pigeon would be unlikely to derive useful information from the model.

Geomagnetic field intensity at Hawaii for the last 420 kyr from the Hawaii Scientific Drilling Project core, Big Island, Hawaii

Four hundred twenty five new paleointensity (Thellier‐Thellier) determinations (out of 545 analyzed samples) have been obtained from core HSDP, which penetrates about 1000 meters (208 flows) of the Mauna Loa and Mauna Kea volcanic series encompassing the last 420 kyr. Rock magnetic investigations identify pseudo‐single‐domain magnetite as the main magnetic mineral. Inclinations are shallower than expected from a geocentric dipole field but are consistent with data from other geographical regions at the same latitude. The inclination record reveals three episodes of negative inclination whose interpolated age correlates well with that of known geomagnetic events. The paleointensity record from the Mauna Loa sequence is not very detailed and does not allow precise comparison with other data in the 0–50 kyr interval. The record from the Mauna Kea sequence, on the contrary, is very detailed and documents relatively short‐lived episodes of low and high field strength from 15 to 60 μT. The average virtual dipole moment (8.7±3.0 1022 A.m2) is not significantly different from the value reported by Kono and Tanaka [1995] for the last 2.5 Myr. A comparison with other data from Hawaii and other geographical regions is described in detail. There are no drastic changes in paleointensity with the inclination anomaly, in agreement with previous results from Hawaii but in contrast with most published results which, however, consider data from polarity transition. Spectral analysis of a particularly detailed portion of the record, between 420 and 326 kyr, documents significant periodicities at 36, 8, 5, and 4 ka in the inclination record but not in the intensity record, suggesting that changes in time of the inclination are to a certain extent independent from those of the intensity.

A record of the Blake Event during the last interglacial paleosol in the western Loess Plateau of China

A high-resolution record of the Blake Event has been obtained from a 40 m well dug in a loess section at Jiuzhoutai in the City of Lanzhou in the western Loess Plateau of China. The paleomagnetic signal is principally carried by magnetite both in loess and paleosols. The Blake Event is located at the boundary between paleosol S1-c of the last interglacial (equivalent to MIS 5e in deep-sea records) and loess L2-2 (MIS 5d), and is characterized by a triple feature consisting of two short reversed intervals separated by a short normal interval. The directional changes are abrupt, as previously revealed in other records, and suggest geocentric dipole field behavior before, after and during the event, but not necessarily during NR or RN transitions. The age and duration of the event are estimated as between 119.97 and 114.47 kyr BP and 5.5 kyr, respectively, from thermoluminescence dating and astronomically tuned climate stratigraphy based on high-resolution magnetic susceptibility.

Geomagnetic field intensity over the last 42,000 years from core SOH‐4, Big Island, Hawaii

A record of the absolute geomagnetic field intensity spanning the last 42 kyr has been obtained from 100 lava flows (187 successful Thellier experiments) recovered in the top 468 m of core from the SOH‐4 well on the Island of Hawaii. Assuming a linear extrusion rate between the present and the 42 kyr date obtained with a refined K/Ar technique, this corresponds to an average of one flow every 420 years. Rock magnetic analysis identifies low‐Ti content and high‐Ti content magnetites in 75% and 25% of the samples, respectively, as the main carriers of magnetization. The low‐Ti magnetites are very stable while some transformations occur upon heating in the other samples, but both groups allow reliable paleointensity determinations. The directional record documents inclinations consistent with a geocentric dipole field for the last 16 kyr and between 38 and 42 kyr. Between 16 and 38 kyr shallower and, in some cases, negative inclinations, were observed. Because of the high extrusion rate and the exceptional percentage of reliable intensity determinations, the record of the geomagnetic field intensity over the last 42 kyr has a resolution comparable to the best sedimentary records. The geomagnetic field intensity at Hawaii appears to have varied between 18 and 79 µT and is characterized by large fluctuation with peak‐to‐peak amplitudes of 20–25 µT. The observed values are quite consistent with other values obtained from Hawaii for the last 31 kyr (Mankinen and Champion, 1993a, b; Tanaka and Kono, 1991; Coe et al., 1978). The mean intensity over the explored time interval, 45 µT, is slightly higher than the present value of the geomagnetic field in Hawaii.

Miocene paleomagnetic results from southeastern Korea

Paleomagnetic samples were collected from 25 sites of Late Oligocene to Middle Miocene age from southeastern South Korea. The mean direction obtained is D = 39.4°, I = 46.2° with α 95 = 9.5°, which is statistically distinguishable from the present axial geocentric dipole field direction. When combined with the previous reported results of eight sites from this region, we get a mean direction ( D = 40.1°, I = 47.4°, α 95 = 7.9°) for 33 sites of Late Oligocene to Middle Miocene age. This direction yields an early Neogene virtual geomagnetic pole of 56.5°N, 215.2°E with an α 95 of 7.9°, discordant from the Cretaceous reference pole for Korea. The paleomagnetic data suggest that portions of southeastern Korea underwent a large clockwise rotation of about 40° during the Neogene. Two large NE-trending right-lateral faults (the Yangsan Fault and the Dongnae Fault, Fig. 1) cut the study areas. Both these faults show evidence for recent activity. In addition, minor faults with similar trends to these major faults occur. The observed deflections in paleomagnetic declinations were probably caused by rigid block rotations of crustal blocks in a right lateral shear system during the Neogene. The paleomagnetic directions for Miocene aged rocks from southeastern Korea and southwestern Japan are similar in direction and magnitude. This suggests that the cause of the rotations in both areas may be related to a similar tectonic event in this area during the Neogene.

A revised magnetostratigraphy for the Mid-Proterozoic Gardar lava succession, South Greenland

New palaeomagnetic results from the ca. 1300 Ma Gardar lava succession within the Eriksfjord Formation of South Greenland are presented. This study incorporates stepwise thermal demagnetisation of 256 samples from 66 lava flows from all three lava groups in this formation, accompanied by principal component analysis. The results identify two complete reversals (N → R → N) in the upper lava group and two (R → N and N → R) in the lower group, with a possible palaeoexcursion in the lower group. A minimum of two further reversals are predicted in the time interval represented by the sandstone units intercalated with the lava flows. Both complete reversals in the upper group exhibit inclination asymmetries of 30°. This degree of asymmetry is satisfied by a geomagnetic field model equivalent to a dipole axially offset by 570 km. However, a more complex source is not precluded by the present evidence. Four new pole positions are calculated for the Gardar lavas, two for the upper lava group and one each for the middle and lower groups. The middle lava group shows an anomalous mean direction and pole position, which is unrepresentative of the remainder of the lava succession and is thought to record a virtual geomagnetic pole representative of a transitional field direction. The presence of a non-axial geocentric dipole field means that the use of the Laurentian apparent polar wander path for dating the new poles is not valid.

Origin and Modification of Magnetic Fabric in Fine-Grained Detrital Sediment by Depositional and Postdepositional Processes

Glaciolacustrine varved clay of late Wisconsinan age in western New York has stable remanent magnetization and anisotropic magnetic susceptibility (AMS). The results of rock magnetic tests demonstrate that remanence is carried by interacting single domain grains of magnetite, but coarse multidomain grains of magnetite are also present. As in many fine-grained detrital sediments, remanent inclination is anomalously shallow, given the latitude of the area of deposition and the existence of a geocentric dipole field at the time of deposition. The AMS consists of a foliation that is gently inclined to bedding and a weaker lineation in the plane of foliation. Independence of magnetic fabric and direction of remanence can be demonstrated by comparing remanence and AMS at closely spaced sites in the same beds and also between widely spaced sites of different age. The magnetic fabric suggests that deposition is accompanied by transient density currents which align large multidomain grains; magnetic alignment of single domain grains occurs later in a dilute slurry at rest on the surface of deposition, rather than during settling in still water. Compaction after deposition results in decrease of remanent inclination and increase in magnetic foliation. Remanent declination and magnetic lineation are unaffected by compaction.

Paleomagnetism of Quaternary volcanics in the Izu peninsula and adjacent areas, Japan, and its tectonic significance.

A paleomagnetic investigation was made on the Quaternary volcanics in the Izu Peninsula and the adjacent Ashitaka and Izu-Oshima areas located on the northern tip of the Philippine Sea plate, central Japan. This study evaluates the deformation associated with the collision of the Izu block with Honshu. About 900 volcanic rock samples ranging in age from 0.2 to 1.5Ma were measured. Alternating field demagnetization indicates that these samples have stable natural remanent magnetizations. Fine grained titanomagnetite was identified as the carrier of the remanent magnetizations by thermomagnetic analysis and microscopic observation. The paleomagnetic directions from both Ashitaka Volcano and Izu-Oshima island, which are located to the north and to the east of the Izu Peninsula, respectively, are not significantly different from that of the present axial geocentric dipole field. In contrast, the mean directions from the northwestern Izu Peninsula show on average 10° clockwise deflections, whereas the directions from the northeastern to southern Izu Peninsula show 8° to 25° counterclockwise deflections. These deflections of paleomagnetic directions are probably caused by tectonic deformations since 1.5Ma. We propose a simple tilting model to explain these paleomagnetic directions.

Extension tectonics: The Neogene opening of the north–south trending basins of central Thailand

Paleomagnetic samples were collected from late Neogene basalt flows from Thailand. All of these flows are horizontal and are relatively unaltered in thin section. These rocks possess a stable magnetization which is believed to be primary. Samples from 48 lava flows were collected from sites located within the Khorat Plateau, the Chao Phraya‐Phitsanulok Basin, and the mountainous terrane west of the Chao Phraya‐Phitsanulok Basin. These data were combined with previously reported late Neogene data from five flows from western Thailand. Although the average inclination from the 53 sites is indistiguishable from the expected dipole inclination, the average declination has a net clockwise rotation of 13.5±5.8 from the geocentric dipole field. Furthermore, the mean declination values from the 29 flows from the Khorat Plateau are indistinguishable from the present dipole field direction (Dm = 4.3°±7.5°) and indistinguishable from the mean declination from 28 late Neogene volcanic flows from Vietnam. In contrast, the mean declinations from 24 flows collected from central and western Thailand are deflected significantly clockwise (Dm = 24.4°±7.7°) from the geocentric dipole field direction. The differential rotation between western and central Thailand versus the Khorat Plateau suggests that Indochina is composed of at least two structural blocks which underwent a different rotational history. These observations, when combined with geologic and geophysical data from the Chao Phraya‐Phitsanulok Basin, Gulf of Thailand, and the intermontane basins of western Thailand, suggest that the rotations are recording a late Neogene phase of E–W extension of these basins. We suggest that the formation of these basins and the related basaltic volcanism developed in reponse to subduction of the Indian plate under western Burma. We envision the tectonics of this region is similar in style to the Basin and Range region of the western United States. Last, we have observed field relationships from some of the rhyolites located in the central basin. Although these rhyolites are reported to be Mesozoic or Paleozoic in age, our field observations and a K‐Ar age date show that at least some of these rhyolites are younger than the basalts. We suggest that the rhyolites form a bimodal suite with the basaltic rocks which were erupted in the later stages of the extension.

Magnetism of igneous rocks from the Tourmakeady and Glensaul inliers, W. Ireland: mode of emplacement and aspects of the Ordovician field pattern

SUMMARY Results of palaeomagnetic and rock magnetic studies are reported mainly for quartz porphyry collected at 18 sites from an Ordovician igneous-sedimentary succession at Tourmakeady and Glensaul in Co. Mayo, Eire. Both magnetite and haematite are shown to be prominent magnetic constituents. The magnetite is usually in the form of large multidomain grains; these are held responsible for a soft remanence component (called the B-magnetization) with D, I= 000, +75; α95= 3.7°, i.e. practically identical to the direction of the present axial geocentric dipole field in the area. Hence, it is inferred that the B-magnetization, which is dominant at most sites, is of recent origin. Haematite, which has a more patchy occurrence, apparently formed through high-temperature oxidation of magnetite. Haematite is the main carrier of the palaeomagnetically significant component, the A-magnetization, which gave well-grouped in situ directions (29 samples from 10 sites) with a mean value of D, I= 182°, +6°, corresponding to a north pole at 31° N, 169° E; dp, dm = 2.1°, 4.2°. Correction for folding along the Glensaul-Tourmakeady are shows that the A-magnetization is of post-folding age. An earlier suggestion that the rocks represent a Lower Ordovician extrusive suite is discussed and discounted. It is inferred that the quartz porphyries constitute syntectonic intrusions having formed during the Taconic orogeny. These rocks as well as their characteristic magnetization are therefore regarded as late Ordovician in age. It is concluded that a c. 50° relative clockwise shift between the mid-Ordovician and late Ordovician-Silurian geomagnetic field axes, previously demonstrated for Scotland, has now been recognized also in Ireland.

Differential Rotation of Northeastern Part of Southwest Japan

A paleomagnetic study has been carried out on the Miocene volcanic and sedimentary rocks from the Yatsuo area, the northern part of the Chubu district. Stable primary magnetizations are recognized from 8 sites among 36 sites ranging from 17 to 7Ma in age. Eight sites are grouped into two by their declinations; the lower 4 sites show easterly deflected declinations (mean direction: D=12.6°, 1=47.1°, α95=7.0°) and the upper 4 sites show no significant deflection from the axial geocentric dipole field (mean direction: D=-2.5°, I=55.6°, α95=6.6°). This result suggests that the Yatsuo area was subjected to a clockwise rotation about 13° between 15 and 12Ma, possibly synchronous with the rotation of Southwest Japan. The smaller rotation angle than that estimated for Southwest Japan (47°) suggests the differential rotation between the block which contains the Yatsuo area and the main part of Southwest Japan.

A New Palaeomagnetic Result from the Lower Cretaceous of East-Central Africa

Summary The mean direction of natural remanent magnetization (NRM) in eight samples of syenite of Lower Cretaceous age from the Mlanje Massif, Malawi, has declination 333 and inclination - 54 (upward) after magnetic cleaning. The stability of remanence in alternating magnetic fields suggests that the NRM is primary. On the assumption of a geocentric dipole field, the palaeomagnetic pole is calculated at 60 N, 98 W. This is in agreement with previous results from volcanics of the Lupata Series nearby in Mozambique, which are of similar age. Combination of results from the Mlanje Massif and the Lupata volcanics yields an estimate of the palaeomagnetic pole for the lower Cretaceous of this part of Africa at 609 N, 993 W, with circular standard deviation (related to palaeosecular variation) of 10.0. Directions of natural remanent magnetization in eight samples from the Lauderdale erosion crater (16 1' S, 35 329 E) on the southern side of the Mlanje Massif in south-eastern Malawi are reported. The massif is a syenite-granite complex and has been mapped by Garson & Walshaw (in preparation). The samples were collected from the outer syenite ring of the complex, along about 150 metres of a stream running at right-angles to its contact with the Palaeozoic basement. The boundary of the complex is transitional, the adjacent basement having been syenitized. Only the intrusive body has been sampled. It is several kilometres in diameter, and it is likely that the section which has been sampled represents a long period in the cooling history of the body. Therefore, in spite of the small collection there is reason to believe that palaeosecular variation may have been effectively averaged out in calculating the mean direction, and that the dispersion may be a fair estimate of the palaeosecular variation. Two recent K-Ar age determinations by Snelling (1966) indicate the Cretaceous age of the complex : biotite from perthosite (sample WllOS) from the outer ring at Chambe indicated 1 16T 6 m.y., and biotite-hornblende mixture from quartz-syenite (sample W1112) from the Nanchidwa Hills, some 19 km east of the Lauderdale Crater gave 128 T 6 m.y. Gough & Opdyke (1963) have reported the palaeomagnetism of the Lupata alkaline volcanics and also of one site in redbeds of the Lupata Series, some 160 km south-west of Mlanje, in Mozambique. McDougall made two determinations on anorthoclases from the Lupata volcanics at 110.5 and 106.7m.y. (Gough et al. 1964). Because Gough & Opdyke sampled seven sites in single lava flows, and hence sampled no more

DÉSAIMANTATION THERMIQUE ET ANALYSE STATISTIQUE DES DIRECTIONS DE SÉDIMENTS CARBONIFÈRES ET PERMIENS DE L'EST DU CANADA

Thermal demagnetization of red beds (109 specimens, 42 sites) from the Maritimes and Gaspé shows that the fossil magnetization consists of a large, thermally stable component and of small, unstable components that disappear at about 400 °C. Thermal cleaning of the whole collection was therefore carried out at 450 °C to demagnetize these soft components. This procedure changes the mean declination and inclination by 6° and 7° respectively and also changes the distribution of directions about the mean. A statistical analysis shows that the site mean directions (after demagnetization) obey a Fisher's (1953) distribution whereas the individual specimen directions do not. This result is of interest in two connections: first, it shows that, in this instance, the use of Fisher's (1953) statistics to analyse the results is not permissible when unit weight is given to the specimen, but that their use is legitimate when unit weight is given to the site; second, the result is consistent with the hypothesis that the earth's field, when sampled in this way (paleomagneticalty), can be represented by an axial geocentric dipole field perturbed by randomly directed magnetic components at the earth's surface. Comparison with other results available for this geological period shows a certain divergence among the paleomagnetic poles for North America. It is suggested that a partial demagnetization of these earlier collections might reduce the observed divergence.

Magnetism in Heard Island rocks

The directions of remanent magnetization in igneous rocks from Heard Island show both negative and positive polarities. Their mean, irrespective of sign, is close to the hypothetical axial geocentric dipole field but differs from it by an amount which is about equal to the error in the determination. This divergence may not be real and requires further study, but if it is real it could be explained by a southerly movement of Heard Island during the Cenozoic as suggested by J. T. Wilson's hypothesis of earth behavior. More detailed sampling at Heard and also in the Kerguelen archipelago could provide a critical test for this hypothesis. The dispersion of directions is measured by a circular standard deviation of 11°, which is much less than the secular variation in these southern latitudes today.

The paleomagnetism of the Shawa Ijolite

Oriented cores have been drilled at seven sites in ijolite exposures of the Shawa carbonatite complex, Southern Rhodesia. The ijolite has been dated at 209±16 m.y. from Rb → Sr decay. The directions of natural remanent magnetization are not consistent between sites. The ijolite proved difficult to treat by partial demagnetization in alternating fields, as instability of the direction of remanent magnetization tended to develop with demagnetizing fields only slightly larger than those required to remove isothermal magnetizations. Nevertheless, consistent directions of remanent magnetization were secured from five of the seven sites, after alternating-field treatment. On the assumption of a geocentric dipole field, the mean south magnetic pole position is 64.2°S, 85.6°E, with radius of 95 per cent confidence 14.1°. This is in excellent agreement with other Mesozoic pole positions from Africa.

The significance of paleomagnetic results from Africa

Paleomagnetic results from Africa are critically reviewed and polar-wander curves are presented for the early Precambrian and for the interval from the Permian to the present. The data are shown to be consistent with the hypothesis that the dispersion of the geomagnetic field with respect to an axial geocentric dipole field was similar to its present dispersion both in the early Precambrian and in Mesozoic times. Strong evidence is presented that relative movement between Africa and the pole was slight during the Mesozoic. Similar evidence exists for Australia. Mesozoic paleomagnetic poles from the four southern continents are shown to form widely separated groups. Polar wander alone cannot account for both the divergence of the poles and the stability of the Mesozoic paleomagnetic poles relative to Africa and Australia. The results can be reconciled, however, by the supposition that relative movement has occurred between the southern continents since the Mesozoic. A Mesozoic reconstruction of Gondwanaland and also a tentative reconstruction for the Permian are proposed. These are based on the most reliable paleomagnetic results available.