Introduction

Most science is comparative. Earth scientists often need to know if a particular phenomenon has had an effect, or if there are differences in a particular variable measured at several different locations. For example, what is the permeability of sandstone with and without carbonate impurities? How does turbidity vary across a glacial lake? How well does the distribution of dew point temperature predict rainfall? But when you make these sorts of comparisons, any differences among areas sampled or manipulative experimental treatments may be real or they may simply be the sort of variation that occurs by chance among samples from the same population.

Here is an example of commercial importance. Most diamonds are mined from kimberlite deposits, which are volcanoes that have risen from great depths in the Earth's mantle at high speed. Sometimes, the kimberlite brings along diamonds that have formed at high pressures and temperatures. But not all kimberlites contain diamonds, and finding them within these rocks is quite difficult.

Fortunately, many kimberlites contain large amounts of the mineral garnet. A prospector noted that the garnets present in diamond-rich kimberlites were slightly darker than those in kimberlites lacking diamonds, and subsequent research suggested that the change in color was caused by the presence of small amounts of oxidized Fe, or Fe3+. To test if oxidized garnets could be used to predict the presence of diamonds, the prospector collected 14 garnet samples: seven from diamond-bearing deposits of kimberlite, and seven from kimberlite without diamonds (Table 6.1), and measured their Fe3+ content.