Previously, I wrote about some of the challenges of studying the mantle. I also wrote about mass spectrometers—this was not accidental, as they were used heavily in the research discussed here. If you have not read those items already, you should do so before continuing. Also, if you are not familiar with isotopes, you may wish to get more familiar with those as well.
Although Big Ben is the dominant feature on Heard Island (seen above with a bow wave and some poorly-defined Von Karman vortices), there is a smaller volcanic edifice, Mt. Dixon, on the Laurens Peninsula (to the NW, right in the bow wave from Big Ben). Mt. Dixon is home to many lava flows, which can be seen on Google Earth, and are believed to be as young as 200 years or less.
The major-element composition (Si, K, Na) of the lavas from Big Ben and Mt. Dixon can be quite different. Big Ben generally has basalt and trachybasalt composition (low SiO2, moderate K2O + Na2O), while the Mt. Dixon and the other cones on the Laurens Peninsula show a much wider range, from basanite to trachyte (wide range of SiO2, generally higher K2O + Na2O).
Where things really get interesting is in looking at the isotopes. Specifically, Barling et al. looked at the isotopes of Sr, Nd, and Pb isotopes.[2,3] Some of those isotopes (86Sr, 144Nd, and 204Pb) are stable and non-radiogenic. That is, they do not decay away, nor are they formed from radioactive decay. The other isotopes studied (87Sr, 143Nd, 206Pb, and 207Pb) all are stable, but are the products of radioactive decay (87Rb, 147Sm, 238U, and 235U, respectively).
The ratio of radiogenic/non-radiogenic isotopes can be used to identify different sources, sort of like fingerprinting. To get high concentrations of radiogenic isotopes means that the rock’s history includes lots of the radioactive parent. Low concentrations of radiogenic isotopes means that the source rock has relatively little of the radioactive parent.
This is important, because although isotopes of an element are chemically similar, different elements behave differently from a chemical standpoint. Some are more often found in the crust than the mantle, while others are the opposite, depending on the compatibility of the element in mantle minerals.* Uranium is generally incompatible, and preferentially moves into the continental crust. Crustal rocks, would be likely to have a high ratio of radiogenic to non-radiogenic lead (product of uranium decay). Mantle rocks would have a lower ratio of 206Pb/204Pb, and similarly for 207Pb/204Pb.
Zindler and Hart (1986) proposed that oceanic basalts can be treated as mixtures of four components, each having a distinct chemical (and isotopic) composition.[4, via 2] Barling and Goldstein found that the Heard Island lavas exhibit a range of compositions consistent with mixing between two sources. Neither of those sources matches the compositions suggested by Zindler and Hart. For the first Heard Island source, three explanations are given why that may be the case:
- The Heard Island source is a mixture of two Zindler and Hart sources
- That same Heard Island source is a fifth distinct mantle source
- It’s more complicated; the two Zindler and Hart sources in question define a spectrum, and the Heard Island source lies along that spectrum
Barling and Goldstein (1990) favored case 3, which they argue is reasonable given that recycling continental crust is likely to give a wide range of isotopic compositions.
Barling et al. (1994) built off of the results presented by Barling and Goldstein (1990), and focused on two main questions:
First, what is the origin of continental crustal signatures in oceanic basalts; are they inherited from the mantle source region, or are they caused by shallow contamination? If they originate in the mantle, how much continental material is present, how is it distributed and in what form, and how and when did it become incorporated into the mantle? Second, what are the origin and timing of enrichment of the sub-Indian Ocean mantle?
Perhaps some clarification is needed about what is at issue. Since it is clear there is some continental influence expressed by the Heard Island lavas, where in the history of that magma did mixing with continental crust occur? Was there a chunk of intact continental material relatively near the surface which partially melted as the basalt came upward through it? Or was there continental material which has been mixed in to the mantle beneath the Indian Ocean? If that occurred, when, and under what conditions?
Their data, and particularly the lead isotopic data (207Pb, 206Pb, and 204Pb), lead them (pardon the pun) to conclude that the component with a high-87Sr/86Sr is derived from marine (ocean) sediments subducted into the mantle at least 600 Ma before present, and probably 1–2 Ga. Modeling of the Sr isotope ratios and total concentrations, along with thermodynamic considerations, suggest that partial melting followed by partial crystallization from the magma is unlikely. That is, recycled crustal material is needed to make things work.
Barling et al. (1994) found that the overall isotopic compositions of the lavas suggest, if crustal material is indeed being recycled into the mantle, the subduction occurred around 1–2 Ga. That timing makes it far too early to be related to subduction beneath the paleo-supercontinent Gondwana.
Finally, the paper closes with the suggestion that, although Heard Island and Kerguelen Island are separated by 440 km, the two may be manifestations of the same plume head and hotspot. They note that the distance between the islands is quite small for separate hotspots, yet is obviously large for being just one hotspot. Perhaps the 2015 Heard Island expedition can collect samples which will give insight into resolving this question.
 Barling, J.; Goldstein, S. L. (1990) Extreme isotopic variations in Heard Island lavas and the nature of mantle reservoirs. Nature 348:59-62, doi 10.1038/348059a0.
 Barling, J.; Goldstein, S. L.; Nicholls, I. A. (1994) Geochemistry of Heard Island (Southern Indian Ocean): Characterization of an Enriched Mantle Component and Implications for Enrichment of the Sub-Indian Ocean Mantle. Journal of Petrology 35:1017-1053, doi 10.1093/petrology/35.4.1017.
 Zindler, A.; Hart, S. (1986) Chemical Geodynamics. Annual Review of Earth and Planetary Sciences 14:493-571, doi 10.1146/annurev.ea.14.050186.002425.
* This turns out to be crucial for things like uranium-lead dating, where the mineral zircon generally crystallizes with 10-1000 ppm U, but does not incorporate Pb. All the Pb found in a zircon can be assumed to come from uranium decay or laboratory contamination (which has a known isotopic composition).