The mantle can be sampled directly only very rarely. Geochemists have thus come to rely heavily on mantle-derived magmas to study the composition and evolution of the mantle. Only those compositional features that are unaffected by magmatic processes are useful as tracers of mantle processes. These include radiogenic isotope ratios such as those of He, Sr, Hf, and Os, stable isotope ratios, and ratios of highly incompatible elements or elements of similar incompatibility, such as Ba/Nb or Pb/Ce. The term ``incompatible'' denotes a preference of the element for a melt over mantle minerals. Highly incompatible elements will partition entirely into the melt under most circumstances, so that the ratio of two such elements in a basalt will be virtually identical to that ratio in its source. This is also true to a lesser degree of ratios such as La/Sm and Zr/Nb, as Zr and Sm are not highly incompatible elements.
Of these mantle tracers, isotope ratios have proven the most useful
because they are insensitive to magmatic processes and because they are
functions of time. Thus the observation that two basalts have different
Nd/
Nd ratios demonstrates not only that their mantle
sources had different Sm/Nd ratios, but also that these differences
existed for geologically long times (10
years or more).
From a geochemical perspective, the mantle can be divided into 3 parts. The first is the mantle lithosphere. Most of the oceanic lithosphere is probably chemically similar to the depleted mantle, but the continental lithosphere is more heterogeneous. The second is the source of mid-ocean ridge basalts (MORB), often referred to as ``depleted mantle.'' It is characterized by incompatible element depletion and radiogenic isotope ratios indicating that this depletion must have occurred billions of years ago. The depleted mantle is more homogeneous than other mantle or crustal reservoirs. Most mantle geochemists and geophysicists believe this reservoir occupies the upper mantle beneath the lithosphere. There are some dissenters, most notably D. L. Anderson [e.g., Anderson et al. 1992], who places the depleted mantle within the seismic transition zone. The final part is the source of mantle plumes. Plumes must originate from some thermal boundary layer, such as the core-mantle boundary or the 660 km discontinuity.
In the following sections, I discuss the advances made in the last 4 years in understanding the evolution of these 3 parts of the mantle. I then discuss some significant advances in Re-Os isotope and noble gas isotope geochemistry. Finally, I discuss the identification of a common component in mantle plumes. Due to space limitations, this review focuses on large scale mantle processes, and does not consider U-series isotopic studies of magmatic processes or subduction zone processes. The reader is referred to Gill [1992] and Hawkesworth [1993], respectively, for reviews on these subjects.