Sparse matrix screening involves a combination of conditions (varying: pH, buffer, additive and precipitant) that have previously generated protein crystals.

This type of screening process is often recommended as the first method to attempt with a protein that has not been previously crystallized. Jancarik & Kim introduced this type of screening 1991.

Benefits:
Commercially available (see below)

Drawbacks:
Biased toward known crystallization conditions
Difficult to make a statistical conclusions due to ‘randomness’ of sampling

Here is a spreadsheet of the overlap between a number of commercially available screens that was adapted from UCLA. Sparse matrix screening is also utilized by Microlytic and Molecular Dimensions in their commercial products.