The ca. 570 Ma Catoctin volcanics, exposed in the Blue Ridge of northern Virginia, include metamorphosed rift-related basalts extruded during breakup of the supercontinent Rodinia. Field relationships, petrography, and geochemistry are used to decipher the stratigraphy for two areas of the volcanics, one at the base of the formation, and the other near its top. Geochemical characteristics of sequential flows can be explained by fractional crystallization of minerals commonly occurring in basalts. Intervening flows with slightly different geochemical features that cannot be explained by fractional crystallization from magma corresponding to an underlying flow, or by crustal contamination, most likely represent new pulses of magma. Positive Pb and negative Nb anomalies, coupled with ε Nd values of +1.5 to +4.5 and model ages that exceed crystallization ages by over 500 m.y., suggest that an enriched component was added to the mantle prior to melting. This component resulted from devolatilization of subducting sediments prior to assembly of Rodinia, likely during Grenville orogenic events. Stratigraphic control shows that lower lavas in the sequence contain more of the enriched component than upper flows.

Ocean island basalt (OIB) and OIB-like basalt are widespread in oceanic and continental settings and, contrary to popular belief, most occur in situations where mantle plumes cannot provide a plausible explanation. They are readily distinguished from normal mid-ocean ridge basalt (N-MORB) through ΔNb, a parameter that expresses the deviation from a reference line (ΔNb = 0) separating parallel Icelandic and N-MORB arrays on a logarithmic plot of Nb/Y versus Zr/Y. Icelandic basalts provide a useful reference set because (1) they are by definition both enriched mid-ocean ridge basalt (E-MORB) and OIB, and (2) they represent a larger range of mantle melt fractions than do intraplate OIBs. Virtually all N-MORB has ΔNb < 0, whereas all Icelandic basalts have ΔNb > 0. E-MORB with ΔNb > 0 is abundant on other sections of ridge, notably in the south Atlantic and south Indian oceans. E-MORB and N-MORB from this region form strongly bimodal populations in ΔNb, separated at ΔNb = 0, suggesting that mixing between their respective mantle sources is very limited. Most OIBs and basalts from many small seamounts, especially those formed on old lithosphere, also have ΔNb > 0. HIMU OIB (OIB with high 206 Pb/ 204 Pb values and therefore a high-µ [U/Pb] source) has higher ΔNb on average than does EM (enriched mantle) OIB, consistent with the presence of recycled continental crust (which has ΔNb < 0) in the EM source. Although EM OIBs tend to have the lowest values, most still have ΔNb > 0, suggesting that a relatively Nb-rich component (probably subducted ocean crust) is present in all OIB sources. The OIB source components seem to be present on all scales, from small streaks or blobs of enriched material (with positive ΔNb) carried in the upper-mantle convective flow and responsible for small ocean islands, some seamounts, and most E-MORB, to large mantle upwellings (plumes), inferred to be present beneath Hawaii, Iceland, Réunion, and Galápagos. It is not possible to identify a point on this continuum at which mantle plumes (if they exist) become involved, and it follows that OIB cannot be a diagnostic feature of plumes. The geochemical similarity of allegedly plume-related OIB and manifestly nonplume OIB is the first part of the OIB paradox. Continental intraplate transitional and alkali basalt in both rift and nonrift (e.g., Cameroon line) settings usually has positive ΔNb and is geochemically indistinguishable from OIB. Continental volcanic rift systems erupt OIB-like basalt, irrespective of whether they are apparently plume-driven (e.g., East Africa, Basin and Range), passive (e.g., Scottish Midland Valley) or somewhere between (e.g., North Sea basin). Magma erupted in passive rifts must have its source in the upper mantle, and yet it is always OIB-like. N-MORB–like magma is only erupted when rifting progresses to continental break-up and the onset of seafloor spreading. Continental OIB-like magma is frequently erupted almost continuously in the same place on a moving lithospheric plate for tens of millions of years, suggesting that its source is coupled in some way to the plate, and yet the Cameroon line (where continental and oceanic basalts are geochemically indistinguishable) suggests that the source is sub-lithospheric. The causes and sources of continental OIB-like magma remain enigmatic and form the second part of the OIB paradox.

ΔNb is a geochemical construct of Fitton et al. (1997) based on two trace-element ratios, Nb/Y and Zr/Y, plotted against one another on a log-log diagram. ΔNb is defined in relation to a diagonal line on the diagram separating Nb-enriched Icelandic tholeiite from Nb-depleted mid-ocean ridge basalt (MORB). Neither crystallization differentiation nor partial melting of a peridotite mantle source can drive compositions below the line, with negative ΔNb, to locations above the line, with positive ΔNb, or vice versa. The parameter ΔNb thus is taken to indicate distinct mantle sources. Crystallization differentiation along the East Pacific Rise has almost no effect on ΔNb. The low and similar partition coefficients for Y, Zr, and Nb during shallow crystallization differentiation ensure that data points for primitive and evolved basalt as well as rare iron-enriched andesite and dacite plot virtually atop one another on the diagram, with all differentiates retaining negative ΔNb. Only enriched MORB (E-MORB) obtained from both ridge segments and the summits of some near-ridge seamounts, which in the extreme includes alkalic olivine basalt, has positive ΔNb. Normal MORB and E-MORB together produce a curving array that crosses the line on the diagram, and this crossing trend indicates heterogeneity in the source. At Iceland, on the other hand, tholeiitic ferrobasalt and abundant andesite, dacite, and rhyolite are widely separated from primitive basalt on such a diagram, plotting at successively higher Nb/Y and Zr/Y at positive ΔNb. The most enriched rocks at Iceland are rhyolites with approximately the geochemistry of the inferred common mantle component, FOZO (focal zone). These rocks cannot be related to each other by either crystallization differentiation or partial melting of a peridotite source. Calculations show that mixing of only 1–2% rhyolite or ∼5% evolved basalt with primitive basalt can shift Nb/Y, Zr/Y, and in certain circumstances ΔNb to higher values; it can also shift chondrite-normalized patterns of rare earth elements in primitive basalt from negative to positive slopes, a general attribute of Icelandic lava suites and the progression along Reykjanes Ridge approaching Iceland from the south. Both isotopic ratios and other trace-element concentrations and ratios are consistent with such mixing. At Iceland, the usual differentiated and enriched mixing component is ferrobasalt with TiO 2 contents so high that it cannot be derived from commonly construed mantle peridotite; a Ti phase such as ilmenite, titanomagnetite, or—at high pressure—rutile, must be in the source. Most picrites, on the other hand, have lower amounts of Nb, Y, Zr, and TiO 2 than primitive MORB. These attributes have led previous workers to propose a source in recycled ocean crust for both types of basalt, brought to the vicinity of Iceland in a mantle plume. However, given the generally enriched character of all Icelandic basalt and the association with rhyolite, I suggest instead that they all are linked to a common source and that they represent successive fractional melts of gabbroic cumulates originally produced during differentiation of ancient granitic continental crust. Much later, during the early stages of the current rifting episode in the North Atlantic, the dense gabbro sank into the warming upper mantle by means of convective instability, converting to eclogite. As Greenland separated from Europe, the eclogite experienced decompression partial melting that contributed to the eruption of flood basalt, and some of it is retained in the Iceland melt source to this day. The high ΔNb of the basalt at Iceland is thus less an indication of mantle source heterogeneity than it is of mixing between depleted basalt and petrogenetically unrelated mafic to silicic differentiates with an ultimate origin in ancient continental crust. It is not evidence for the existence of a mantle plume.