Given the growing controversy about whether or not plumes of mantle rock can rise from the core-mantle boundary to source large igneous provinces (see Geoscience consensus challenged in EPN January 2004) the hypothesis has been tested by seeking material in hot-spot lavas that may have crossed from the outer core into the deepest mantle.  Some hot-spot lavas contain traces of Osmium-186 that may have formed by decay of an unstable platinum isotope (¹⁹⁰Pt) that is most likely to be enriched in the core, thereby supporting the hypothesis.  Another isotopic approach is to look at tungsten isotopes (Scherstén, A. et al. 2004.  Tungsten isotope evidence that mantle plumes contain no contributions from the Earth’s core.  Nature, v. 427, p. 234-237).  Tungsten, like osmium, has a strong affinity for iron, and the bulk of terrestrial W is likely to be present in the core.  One isotope ¹⁸²W forms from the decay of an unstable isotope of hafnium ¹⁸²Hf, whose half life is geologically short (about 9 Ma).  As a result all ¹⁸²W in the Earth must have been produced in the first 60 Ma of the planet’s evolution.  Moreover, hafnium is likely to favour the mantle far more than the core, so most ¹⁸²W seems likely to be present in the mantle and the core should be depleted in it.  This is borne out by comparing values in primitive meteorites with those in mantle-derived lavas; the mantle is enriched by comparison.  So, if there was significant chemical exchange between the core and mantle a lot of tungsten with very low ¹⁸²W should contaminate the lower mantle.  If plumes did rise from the core-mantle boundary, then lavas derived from them ought to have anomalously low ¹⁸²W contents. Scherstén and colleagues from the University of Bristol and the Australian National University show that Hawaiian lavas (the same samples used to suggest a mantle-wide plume beneath Hawaii using osmium isotopes) and South African kimberlites do not show this signature, and argue convincingly that the osmium data must represent another source of contamination, probably recycled crustal rocks.  However, that does not rule out a plume rising from the core-mantle boundary, just that the core did not play a significant geochemical role.