In this thesis, we take the first steps towards measuring the uctuations in HI emission before reionization which carry information on the first X-ray emitting compact objects and hot interstellar gas heated by the deaths of the first stars (ancient demons). First, we show that existing and planned interferometers are sensitive enough to place interesting constraints on the astrophysics of X-ray heating. Second, we obtain first upper limits on the pre-reionization 21cm fluctuations with the Murchison Widefield Array. We also use these measurements to explore the impact of low-frequency systematics, such as increased foreground brightness and the ionosphere.
We discover that contamination by fine-scale frequency structure introduced by the instrument is the leading obstacle to measuring the 21 cm power spectrum before reionization. This motivates the design of a next-generation experiment, HERA, with acceptable levels of intrinsic spectral structure. We also perform a careful examination of whether traditional calibration strategies are able to remove instrumental spectral structure to an acceptable level. While existing calibration techniques have critical flaws, we discover promising strategies for overcoming these flaws.