A small piece of Earth's upper mantle from ~75 km-depth

Garnet peridotite from the subcontinental lithosphere below southernmost Patagonia, Pali-Aike, Chile, brought to the surface in Pleistocene-Holocenean alkali basaltic magma. Olivine grains (light green) in xenoliths from Pali-Aike exhibit rims that are depleted in hydrogen relative to their central cores. The hydrogen profiles represent a dehydration process having occurred during ascent in the host magma. Experimental hydrogen diffusion data were used to fit the observed profiles, constraining the xenolith’s ascent rate. Some of these xenoliths appear to have reached the surface from 50-70 km depth within several hours, a surprisingly rapid rise for basaltic magma.

Reference

Demouchy, S., S.D. Jacobsen, F. Gaillard, and C.R. Stern (2006) Rapid magma ascent recorded by water diffusion profiles in mantle olivine. Geology 34, 429-432. PDF

This rock represents "fertile" upper mantle with olivine (light-green) orthopyroxene (dark-green or brown) clinopyroxene (Cr-diopside, emerald-green) and garnet (red). The bulk composition is close to Ringwood's "pyrolite". A partial melt of this material would produce basalt.

I made this image (left) on a flatbed scanner.
Photo edge is about 5 cm

Below: Series of polarized (E//a) infrared spectra measured across an olivine grain from Pali-Aike (PA-7). Rims depleted of hydrogen are interpreted to be evidence for dehydration during ascent. Does the mantle contain more water than we thought? This is Figure 3 of Demouchy et al. (2006).