THE FACTS

A major peculiarity of giant dike swarms and associated flood basalts is an enormous supply of magma from an upper mantle source to the Earth's crust. For example, a single dike of the Proterozoic Mackenzie swarm in Canada likely carried a volume of magma equivalent to the 100 years-worth of melt production of the contemporary global mid-ocean ridge system. The large volumes of magma, as well as the high magma fluxes (individual giant dikes may have been emplaced on a time scale of weeks to months, and the entire magmatic episodes might have occurred on a time scale on the order of 10⁵ years) imply a major thermal or compositional (or both) perturbation in the upper mantle, or a perturbation in the lithospheric stress field, over a relatively short time.

The inferred high rates of magma production require a robust advection of hot material from the upper mantle and/or a storage mechanism (ponding, underplating) plus a release mechanism (diking? stress valve?). The delivery of melt from the mantle to the surface may not have been a single-stage process, or a steady-state process. Thermo-mechanical and chemical arguments alone do not constrain the depth from which the hot/fertile material was supplied to the melting area (in or below the thermal boundary layer (TBL) or the base of the lithosphere or in or below the crust). It is not certain that the source region of LIPs is strongly superheated. The smaller the holding tank (magma chamber, underplate, magma pond) and the larger the host rock undercooling with respect to the solidus is, the shorter the lifetime.

The association of giant dike swarms and flood basalts with mantle plumes (originally proposed by Morgan, 1981) is based on the observation that some flood basalt provinces seem to initiate hotspot tracks (e.g., Duncan & Richards, 1991), and on isotopic signatures of flood lavas (Carlson, 1991). However, many flood basalts are not associated with volcanic chains and many volcanic chains observed on the ocean floor have no associated LIPs or obvious plume heads. In some cases, this may be explained by the removal of the initiating LIPs by subduction.

On continents, most large igneous provinces (Continental Flood Basalts – CFB) occur on or along suture zones adjacent to cratons. When a continent splits, there is often (but not always) a transient burst of magmatism. Steady-state mature ridges take some time to form. Magmatic activity associated with rifting is often correlated with pre-existing faults, transforms, fracture zones etc. (Sykes, 1978). The association of CFBs with Archons (cratonic roots of Archean age), suture zones and pre-existing faults is much closer than it is with time-progressive volcanic chains. Some CFB even occur during continental convergence. There is a cause and effect issue here. Does a large, hot, buoyant mass of magma from deep in the mantle cause the split, or does continental breakup result from plate tectonic processes that then allow the egress of accumulated magma? The occurrence of CFBs during convergence (e.g., Deccan, Keweenawan, Siberian, Alpine/European(?)) and during late stages of breakup (the North Atlantic Tectonic Province) suggests that the association of plumes and breakup is not universal and may incorrect.  The sequence usually assumed is first plume head, then extension, then CFB, and lastly breakup of the continent. However, extension, ponding, diking and breakup may precede  LIP eruption, e.g., the formation of the North Atlantic Igneous province signalled the completion of the Atlantic opening, not the initiation (Ed: see also Iceland and North Atlantic pages).

A seemingly radial pattern of giant dike swarms is often attributed to uplifts caused by mantle plume heads (Ernst et al., 1997), although the amplitude of the resulting uplift is insufficient to accommodate large dike swarms (such as the Mackenzie swarm) by simple magmatic fracturing. Given large magma volumes in the source region, and high fluxes during dike emplacement, it is possible that a major fraction of the observed dike thicknesses was facilitated by non-dilatant mechanisms (e.g., via melting of the dike walls). It follows that the extensional stress in the crust after the dike emplacement may have been larger than prior to the dike emplacement. If so, the radial or fanning patterns of large dike swarms might be due to a self-induced stress field. If a large-scale uplift was centered over the magma source, its main role was probably to provide a gravitational driving force for a long-range lateral magma transport. (Ed: See also “Comments on giant dikes” by Richard Ernst).

Large volumes of magma involved in the flood basalt events pose several mechanical problems. If the giant dike swarms were fed from an intermediate storage region in the crust, there is a problem of space for a large (thousands to millions of cubic kilometers) shallow magma chamber. If dikes ascended directly from a mantle source, and spread horizontally upon reaching a level of neutral buoyancy, it is unclear how large volumes of gravitationally unstable melt might have accumulated in the source region. One possibility is that most of the melting takes place while the lithosphere is under horizontal compression, so that the melt cannot not readily escape via dikes.

Regional uplift due to upwelling mantle generates horizontal compression in the lower lithosphere, which may result in magmatic underplating. A subsequent sub-vertical transport of the ponded magmas may be possible if the horizontal compression is relieved, e.g., by thermally activated creep, gravitational collapse, or tectonic extension. A regional uplift is not a pre-requisite for trapping of large volumes of melt at the base of the lithosphere, which may, in principle be accomplished by any process resulting in horizontal compression of the lower lithosphere.

VOLUME IN CONTEXT

The volumes associated with giant dike swarms and LIPs are indeed enormous but must be considered in context. The largest LIP is the Ontong-Java Plateau. If the 20-km-thick ocean crust there was the result of draining an area 3 times larger than the plateau (because of focusing to the apex of the triple junction) and if 20% melting was involved, then only a 30-km thickness of mantle need be involved. Normal mantle geotherms are well above the solidus in the approximate depth interval from 30 to 50 km depth. For the older giant dike swarms and LIPs, the mantle will have been hotter. A basic question then is: Can magma draining by porous flow to a structural or permeability trap in or beneath the plate accumulate for a sufficient length of time to provide 10⁶ km³ of magma in 10⁶ years? Alternatively, can porous flow in the mantle keep pace with the eruption rates?

THE PARAMETERS

Parameters controlling dike intrusion include magma density and viscosity, pressure in the magma, and stress magnitude and orientation in the lithosphere. Yield stress in negligible compared to typical magmatic overpressures, and host rock density affects propagation only insofar as it affects the already-mentioned stress in the lithosphere. The tectono-magmatic history of the area is also important. A long history of ponding (in the lower crust or upper mantle) may be essential to the formation of long diverging dikes. True radial dikes form as blade-like sheets of magma that move laterally in the shallow crust. Simple gravitational stresses on elevated features like Hawaiian volcanoes (Fiske & Jackson, 1971) or neutral buoyancy (Ryan, 1993) can control the depth of lateral dike injection, but either way, the condition of formation of radial dikes is that the difference between the intermediate and the least principal stresses is less than the magma overpressure (the difference between the magma pressure and the dike-perpendicular stress). No matter what, true radial dikes are shallow, and the regions of lateral dike injection rise, staying about the same depth below summits, as volcanoes grow.

Using the Hawaiian example, volcanoes with radial dike configurations must stand in isolation and not nest against one another. The volcanoes must be high enough that the region of dike injection is not influenced by stresses in the ocean crust or upper mantle. If the volcanoes nest, then rift zones with parallel dikes form. Bora Bora in the Society Islands is an example of a high, freestanding volcano with radial dikes. Isolated seamounts with stellate patterns of lateral rifts (Vogt & Smoot, 1984) are another example. These are abundant in the far western Pacific and formed when the Pacific plate was surrounded by ridges and had minimal internal stress differences. Moreover, these are swarms of “normal” dikes, not giant dikes.

Rose diagrams and maps of dikes are needed to see how truly radial any dike swarm may be. Some so-called radial dike swarms are actually sub-parallel to or sub-perpendicular to rifted margins. Unidirectional dikes from Iceland have been proposed as the origin of the “V-shaped” (chevron) bathymetric ridges that flank the Reykjanes ridge. In other places (e.g., Scotland), later-generation dikes cross-cut earlier dikes. In each case, the dikes have been attributed to thermal plumes. (Ed: See also “Comments on giant dikes” by Richard Ernst).

A plume head could act to bulge up the crust and create a broad region in which the intermediate and least principal stresses become equal, even in the absence of a high volcano. More likely, however, is that the stress configuration was that way to begin with. Some diverging dike swarms, however, may have formed without a broad lithospheric uplift. Uplift is not one of the requirements for radiating dikes. For long-distance lateral injection, the main requirements are for a supply of magma at either high pressure or elevated compared to the surroundings. The surface itself need not be elevated, but the magma chamber needs to be higher than the distant dikes which reach the surface (the artesian effect), if diking is active, rather than just passive filling ofrifts. If there is uplift and no topographic high, the dikes will track the level of neutral buoyancy, which presumably is at the same level as the magma chamber, and distant down-rift eruptions will be unlikely.

Mid-continent LIPs are places where large quantities of melt accumulate beneath the lithosphere in blister-like masses that may indeed cause uplift. Some uplift, combined with tectonic stress or lithospheric weakening, may be the condition required before any magma can escape. But when it does, the eruption rate is great and volume of magma enormous.

THE PHYSICS

A steady state between melting and erupting corresponding to typical CFB fluxes (~1 km³/yr) is unlikely over geologic time. At shorter time scales on the order of ~1 million years, such rates of melt production could conceivably be attained given advection of hot, fertile mantle material, as typically envisioned in the plume theory. Problems arise if the timescale for melting is much greater than the timescale for eruptive activity. Ponding is therefore essential for any theory, and if ponding is allowed, a plume is unnecessary.

In the source region of flood basalts, there is obviously a super- (or at least near-) solidus temperature. As melt is created, it will immediately escape upward because it is gravitationally unstable. Even a “regular” melting column (say, below a mid-ocean ridge) does not contain more than a few percent of melt fraction. Given small melt fractions, basalts may migrate upwards via porous flow in a viscous matrix. The larger the melt fraction , the easier melt transport becomes. If/when melt accumulates in larger bodies, upward transport occurs through a much more efficient self-driven magma fracture (dikes). However, the picture changes once magma leaves the source region and approaches the sub-solidus (by some definitions) base of the lithosphere. In the lithosphere, porous flow is impossible (melt freezes and clogs the pores), and must be accomplished via dikes. Dikes must be thicker than a certain critical width (~ 1 m) to avoid freezing. This implies that dikes must drain sizeable pools of magma. Correspondingly, giant dikes must be fed from very large reservoirs. Long-term existence of large reservoirs of melt at the base of the lithosphere (e.g., in the case where horizontal compression prevents diking) is unlikely, if that is where it ponds, because the melt would exchange heat with the lower lithosphere and freeze. Assuming conductive cooling alone, a 1-km-thick melt layer will lose most of its heat on a time scale of 10⁴ years.

However, the thermal boundary layer is not the same as the lithosphere, which is defined by strength or viscosity and not by thermal gradient. Melts can conceivably pond at density and permeability interfaces, including in and below the region of conduction gradient. In some models, the lower part of the thermal boundary layer can be of the order of 1400°C. The lithosphere is defined by its long-term strength. However, for silicate rocks strength is a function of temperature. An upper mantle hotter than 900-1,000°C may not support any significant deviatoric stress, and thus cannot create the compression required to prevent vertical melt propagation. If melt ponds at a density interface, it will not have any tendency for subsequent vertical propagation. Ponding below a permeability barrier is probably the only option, similar to a sub-axial magma lens at fast-spreading mid-ocean ridges.

DO GIANT DIKE SWARMS RADIATE?

Some of the longest Precambrian dike sets that exist in northern Canada both curve and diverge, but diverging is not the same as radiating. Radiating means extending out from a point in all directions, not just one or two sets that spread out and bend along their great lengths. Where are they on other sides of the focus points? And, why are there not radiating dikes around all plume-uplift centers? Giant dikes do not radiate around Iceland, Columbia River, Parana, Siberia, Deccan, Yellowstone, or even Hawaii, although smaller dikes certainly do radiate around some single volcanic centers (such as Ship Rock). The great Precambrian dikes in Canada may have less to do with domal uplift than with extension of the spherical shell of the lithosphere, which over a great distance will not respond to stress like a flat planar surface. (Ed: See also “Comments on giant dikes” by Richard Ernst).

Late Archean – Early Proterozoic giant dike swarms exposed in the Canadian Shield have a variety of geometries (Figure 3). Dike sets may remain sub-parallel within great sweeping curves or branches (the Matachewan dikes), or form sets of roughly parallel dikes (the Fort Frances dikes), or diverge from a hypothetical focus point that they do not actually reach (the Mackenzie and Mistassini dikes). Dikes far apart and with different orientations may be similar in age (Figure 3). There are no “hotspot tracks” – that is, no volcanic chains progress from a focus suggested for any of the giant dike swarms.

One feature we may all agree on is that dikes form from fluid magmas under pressure that follow (or create) propagating fractures, which have orientations controlled by lithospheric stresses.

We suggest that if a set of parallel dikes propagates far enough, the extensional fractures will curve and/or diverge because they must follow the spreading, curving and domain-limited stress fields of the spherical lithosphere. As the Earth's outer shell is pulled apart, whether to form new rifts, drift the plates, or intrude dikes, there is always a pole around which vectors of extension trace small circles. The sub-parallel fractures must converge toward a pole and diverge away from it. This will obviously not be apparent in sets of shorter dikes. This is plate tectonics, not plume tectonics.

PLATE MIGRATION VECTORS AND DIKE PATTERNS

Linear directions followed by dikes are perpendicular to extension, but they are also in planes that contain the direction of maximum compression. Compression is created by plate collisions, but across the plate interiors it is related to drag on the base of the lithosphere by the convecting mantle. Dikes may form in the same directions as a consequence of plate motions, as has been proposed for widespread Early Cretaceous dikes in New England (McHone, 1988).

THE CENTRAL ATLANTIC MAGMATIC PROVINCE

Fundamental observations by May (1971) first suggested a radial pattern of dikes around the nascent rift zone of central Pangaea, where the Atlantic Ocean crust would be initiated in Early Jurassic time. This apparent pattern, with a focus somewhere east of Florida (perhaps the Blake Plateau), has been used and re-used as evidence that a deep-mantle plume produced the tholeiitic dikes and basalts, as well as continental rifts of central Pangaea (e.g., Morgan 1971, 1981, 1983; Ernst et al. 1995; Wilson 1997; Thompson 1998; Ernst, Giant Radiating Dyke Swarms, this website). Other studies have supported a non-plume plate tectonic origin (e.g., Le Pichon & Fox 1971; Sykes 1978; Lameyre et al. 1984; McHone 2000; Hames et al. 2001). Much new work on this LIP, now called the Central Atlantic Magmatic Province (CAMP), has been published in a new monograph by Hames et al. (2002). Many of the papers in that volume do not support a deep-mantle plume model for the CAMP.

Cartoon maps of CAMP dikes reproduced in plume-model papers have greatly exaggerated the radial geometry. This distortion may be partly due to poor-quality maps for some areas of dikes in Africa and South America. In fact, as discussed by McHone (2000), the dikes occur in separate regions with distinct magma types, and in many overlapping trend groups that do not radiate toward a common point. Even the largest dikes within the regional groups do not physically extend to a common focus at the Blake Plateau. In addition, most large dikes tend to be discontinuous along strike, with en echelon patterns or segments that are offset up to several kilometers. The magmas of such dikes must interconnect in mainly vertical to oblique directions. More commonly, swarms of many smaller dikes of the same magma type do not interconnect in any lateral sense with dikes closer to the “plume center.” The compositional correlations among dikes within groups, the trend and compositional differences between groups, and the mainly vertical flow required by dike structures, indicate that each dike group is derived from its own mantle source, and that each group has its own regional tectonic control.

Figure 5(a): The great N-S dike group that diverges from Charleston, South Carolina northward into Virginia. From Hames et al. (2001).

Figure 5(b): Cross-cutting of older high-Mg NW-SE dikes and rift basins in North Carolina. From Olsen et al. (1991).

However, like the giant dikes of northern Canada, some of the very long dike systems in eastern North America do not remain parallel. A great N-S group diverges from Charleston, South Carolina (not a “plume center”) northward into Virginia (Figure 5a). This narrow dike set cross-cuts older high-Mg NW-SE dikes and rift basins in North Carolina (Figure 5b). These juxtaposed N-S and NW-SE dike sets are formed independently and from different magmas (Ragland at al., 1983). The shorter NW-SE dikes in the southern USA are numerous and occur in a very wide belt from Virginia to Alabama (and beyond under the coastal plain), and they remain sub-parallel from one end to the other because they are in separate, short fractures rather than long, single ones.

In northeastern North America, huge but widespread dikes in Canada and New England diverge to the NE and ENE from a focus point east of New Jersey, but that is also not a plume center. The dikes change their trends across the “New England Salient,” which is a bend in terrane suture zones and primary structures of this section of the Appalachian Orogen. In addition, the giant dikes did not form together in a radial generation, but instead decrease systematically in age from the SE toward the NW (Figure 6).

ANISOTROPY OF MAGNETIC SUSCEPTABILITY (AMS) AND HORIZONTAL FLOW

Our last area of contention concerns the magmas of giant dikes, which are proposed to flow horizontally away from plume centers, possibly for hundreds or thousands of kilometers (Ed: See also Giant radiating dyke swarms page). Long-distance horizontal flow is required in this model, because giant dikes are too large for vertical flow upward up from plume heads over such great lengths. The evidence cited for horizontal flow is often based on anisotropy of magnetic susceptibility (AMS), as measured from oriented drill cores. Although their magmas follow propagating trans-crustal fractures, the dikes themselves may not breach the surface along their entire lengths because their magma density is relatively high.

The AMS of dikes can include a principal axis in a horizontal plane, which has been used as definitive evidence for horizontal dike flow (Tarling & Hrouda, 1993). AMS is mainly caused by small magnetite grains, which crystallize late in the cooling history of basalts (probably after magma flow has ceased). The magnetic anisotropy of basalt is strongly controlled by plagioclase laths, around which the magnetite grains collect in layers along the planar feldspar faces. In addition to alignment by the primary magma flow, plagioclase crystals and resultant AMS fabrics may become oriented by other mechanisms. As basaltic magmas crystallize, a 3-D plagioclase network forms (Philpotts & Dickson, 2000), which eventually collapses and flattens if the magma body is large enough (as in many sills and large dikes). A sub-horizontal preferred crystal orientation results, which must affect the observed magnetic anisotropy (usually a very subtle fabric).

Another cause can be back-flow after fluid pressure diminishes in the later stages of dike activity. Magma movement back down into the dike fractures has been observed in volcanic lava pools, and it could re-orient the dike phenocrysts (commonly feldspars) and surrounding magnetite-rich planes away from their initial upward or oblique directions.

Other features such as contact rip-ups, oriented phenocrysts in glassy contact zones, wisps of melted country rock, and elongate contact cusps can show initial dike flow directions, but even these may differ from later, main-phase magma movements. AMS measurements by themselves are not sufficient to demonstrate long-distance horizontal flow.

SUMMARY

1. Huge volumes of tholeiitic magmas were generated in upper-mantle source regions of giant dikes and may have caused regional uplift or domal swells in the crust. Giant dikes propagated fractures from these magma ponds in response to lithospheric tectonics, with no requirement for deep-mantle plumes.
2. Non-radial patterns of fissure dikes are most common for flood basalt provinces, not radiating giant dikes. True radiating dikes are smaller and reflect local stress fields around volcanoes or high-level magma chambers, not regional extension of the lithosphere.
3. Giant Late Archean-Early Proterozoic dikes in Canada are more accurately described as diverging or arcuate, not radiating. None are connected to a progressive volcanic chain.
4. Very long sets of dikes do not remain parallel but instead diverge as they propagate, because extensional stress orientations vary across large distances in the spherical shell of the lithosphere. There may be stress relationships with continental rifting, plate divergence, mantle convection cells, and mid-ocean ridges, but not “hot spot” plumes.
5. Giant dikes are not known beneath large surface rifts, and radiating patterns of rifts on other planets are not related to magmatic intrusions on Earth.
6. As shown by recent GPS motion vectors in southern Asia, terranes can migrate and rotate along recent compressional sutures or fault zones. Concordant dike swarms generated within tectonic domains may have curving to sub-radial patterns because of variable stress domains.
7. The giant dikes of the CAMP are tectonic features clearly associated in geometry and location with continental cratons and rift margins, and they occur in independent but juxtaposed sets that are not radiating.

• Carlson, R.W. 1991. Physical and chemical evidence on the cause and source characteristics of flood basalt volcanism. Austral. J. Earth Sci. 38, 525-544.
• Burchfiel, B.C.. 2004. New Technology; New Geological Challenges. GSA Today 14, 4–9.
• Burke, K., 1976, Development of graben associated with the initial ruptures of the Atlantic Ocean.Tectonophysics, 36, 93-112.
• Burke, K.C. & Wilson, J.T. 1976. Hot spots on the Earth’s surface. Sci. Am. 235, 46-57
• Deckart, K., Féraud, G., Bertrand, H., 1997. Age of Jurassic continental tholeiites of French Guyana/Surinam and Guinea: Implications to the initial opening of the central Atlantic Ocean. Earth Plan. Sci. Lett. 150, 205-220.
• Duncan, R.A. and Richards, M.A., Hotspots, mantle plumes, flood basalts, and true polar wander, Rev. Geophys. 29, 31-50, 1991.
• Ernst, R.E., Head, W.J., Parfitt, E., Grosfils, E., Wilson, L. 1995. Giant radiating dike swarms on Earth and Venus. Earth Sci. Rev. 39,1-58.
• Fiske, R.F, and Jackson, E.D., 1972. Orientation and growth of Hawaiian volcanic rifts: the effect of regional structure and gravitational stresses. Proc. Roy. Soc. (Lond.), A, 329, 299-326.
• Gudmundsson, Agust. 2003. Surface stresses associated with arrested dykes in rift zones. Bull. Volcanology 65, 606-619
• Hames, W.E., Renne, P.R., Ruppel, C. 2000. New evidence for geologically instantaneous emplacement of earliest Jurassic Central Atlantic magmatic province basalts on the North American margin. Geology 28, 859-862.
• Hames, W.E., McHone, J.G., Ruppel, C., Renne, P. (editors). 2002. The Central Atlantic Magmatic Province: Insights from Fragments of Pangea: American Geophysical Union Monograph. 136, 267 pp.
• Hamilton, M.A., Goutier, J., Mathews, W. 2001. U-Pb baddeleyite age for the Paleoproterozoic Lac Esprit dyke swarm, James Bay region, Quebec: Radiogenic age and isotopic studies. Report 14. Geological Survey of Canada, Current Research 2001-F5, 10 p.
• Lameyre, J., Black, R., Bonin, B., Giret, A. 1984. The magmatic provinces of eastern America, West Africa, and Kerguelen: indications for a tectonic control of within-plate magmatism triggered from above and associated processes. Ann. Soc. Geol. Nord CIII, 101-114.
• Le Pichon, X., Fox, P.J. 1971. Marginal offsets, fracture zones, and the early opening of the North Atlantic. J. Geophys. Res. 76, 6294-6308.
• May, P.R.. 1971. Pattern of Triassic-Jurassic diabase dikes around the North Atlantic in the context of the predrift positions of the continents. Geol. Soc. Am. Bull. 82, 1285-1292.
• McHone, J.G., 1988, Tectonic and paleostress patterns of Mesozoic intrusions in eastern North America, in Manspeizer, W.R., ed., Triassic-Jurassic Rifting: Continental Breakup and the Origin of the Atlantic Ocean and Passive Margins, Part B: New York, Elsevier, 607-619.
• McHone, J. G., 2000. Non-plume magmatism and tectonics during the opening of the central Atlantic Ocean: Tectonophysics316, 287-296.
• Morgan, W.J., 1971. Convection plumes in the lower mantle. Nature 230, 42-43.
• Morgan, W.J., 1981, Hotspot tracks and the opening of the Atlantic and Indian oceans, in The Sea, 7, 443-487.
• Morgan, W.J., 1983. Hotspot tracks and the early rifting of the Atlantic. Tectonophysics 94, 123-139.
• Olsen, P.E., Froelich, A.J., Daniels, D.L., Smoot, J.P., Gore, P.J.W. 1991. Rift basins of Early Mesozoic age, in Horton, J.W. and Zullo, V.A. (editors), Geology of the Carolinas. Knoxville, University of Tennessee Press, 142-170.
• Philpotts, A.R., Dickson, L.D., 2000. The formation of plagioclase chains during convective transfer in basaltic magmas. Nature 406, 59-61.
• Ragland, P.C., Hatcher, R D., Jr., Whittington, D., 1983. Juxtaposed Mesozoic diabase dike sets from the Carolinas. Geology 11, 394-399.
• Ryan, M.P., 1988. The mechanics and three-dimensional internal structure of active magmatic systems: Kilauea volcano, Hawaii. J. Geophys. Res. 93, 4213-4248.
• Ryan, M.P., 1993. Neutral buoyancy and the structure of mid-ocean ridge magma reservoirs. J. Geophys. Res. 98, 22,321-22,338.
• Sykes, L.R. 1978. Intraplate seismicity, reactivation of preexisting zones of weakness, alkaline magmatism, and other tectonism post-dating continental fragmentation. Rev. Geophys. Space Phys. 16, 621-688.
• Tarling, D.H., Hrouda, F. 1993. The Magnetic Anisotropy of Rocks. Chapman and Hall, London, 217 pp.
• Thompson, G.A., 1998. Deep mantle plumes and geoscience vision. GSA Today, April, 17-25.
• Vogt, P.R., 1973. Early events in the opening of the North Atlantic. In: Tarling, D.H., Runcorn, S.K. (Editors), Implications of Continental Rift to the Earth Sciences. Academic Press, New York, 693-712.
• Vogt, P. R. and N. C. Smoot, The Geisha guyots: multibeam bathymetry and morphometric interpretation, J. Geophys. Res. 89, 11,085-11,107, 1984.
• Wilson, M., 1997. Thermal evolution of the Central Atlantic passive margins: Continental break-up above a Mesozoic super-plume. J. Geol. Soc. London 154, 491-495.