Methods for obtaining combinatorial and array-based data as a function of temperature are needed in the chemical and biological sciences. It is presently quite difficult to employ temperature as a variable using standard wellplate formats simply because it is very inconvenient to keep each well at a distinct temperature. In microfluidics, however, the situation is very different due to the short length scales involved. In this article, it is shown how a simple linear temperature gradient can be generated across dozens of parallel microfluidic channels simultaneously. This result is exploited to rapidly obtain activation energies from catalytic reactions, melting point transitions from lipid membranes, and fluorescence quantum yield curves from semiconductor nanocrystal probes as a function of temperature. The methods developed here could quite easily be extended to protein crystallization, phase diagram measurements, chemical reaction optimization, or multivariable experiments.