Magmatic Differentiation: So, all of these sources draw on the melting of mantle rocks, which ought to be pretty uniform. Why do we see such compositional variation in magmas?

Fractional melting:

Mantle rocks like peridotite consist of several different minerals, each with its own melting point. As the rock heats, decompresses, or is infused with water, the minerals with the lowest melting point melt first and begin to move away from the source rock, so a magma is always somewhat more felsic than its source. Consider that the magma erupting at mid-ocean ridges has moved maybe a mere ten km from its source, but whereas that source was ultramafic peridotite, the magma is merely basaltic.
Fractional crystallization. Reverse process of fractional melting. The most mafic minerals in a melt (i.e. those with the highest melting point) will be the first to crstallize out, leaving an increasingly felsic magma.

The sequence in which minerals crystallize out of magma was worked out by N. L. Bowen in the early 20th century: Bowen reaction series.

The vast majority of intrusives we see on the continents are felsic, like granite. Fractional crystallization can't account for this. Remember, in general, continental crust is much more felsic than oceanic. As ultramafic magmas encounter the felsic rocks of the continental crust, they cause the most felsic minerals in those felsic rocks (the ones with the lowest melting point) to melt. Thus, felsic material is added to the magma as mafic material is lost to fractional crystalization. The result is that magmas that have passed through thick layers of continental crust represent highly refined concentrations of felsic materials.

  • Size matters:

    The puzzle remains as to why we might have rhyolitic and basaltic magmas erupting adjacent to one another on the continents. A fourth parameter is the size of the magma chamber. A large one may work its way to the surface and, despite its having assimilated continental crust, retain something of its original composition while a smaller one will be significantly altered by passage through the same thickness of crust.
    But the real world is complicated. Consider the New Jersey Palisades

    The stratification refelcts the sequence in which a homogeneous basaltic magma crystallized after being injected into a joint between two layers of sedimentary rock.

    Igneous land forms

    We now turn to what igneous rocks look like when you see them in the real world. To start with, here's a general hint: Viewed in cross section, igneous bodies that cooled underground tend to have columnar jointing - i.e. they are crossed by vertical cracks that give them a columnar appearance. The image below is of Devil's Postpile National Monument in CA, a particularly beautiful example.

    Even in less perfect forms, columnar jointing is a good way to spot igneous rocks at a distance.

    Intrusions and their geometric forms: An intrusion is any plutonic rock formed from magma that has invaded other rock. Different terms are applied to types of plutons depending on their size and geometry.

    Volcanic intrusions: All of these forms are likely to be made up of plutonic rocks, however the smaller, essentially two dimensional ones have a high surface area to volume ratio. These are likely to show a chill zone at their contact with surrounding rock, which will be volcanic because it chills quickly. If a dike or sill is thin enough, the chill zone is all there is, resulting in a fully volcanic dike or sill.

    Volcanic deposits:

    Additional Information.

    Key concepts and vocabulary: