<p>Optical interferometry is an observational technique that provides the highest spatial resolutions available in the optical. By interfering light from separate telescopes, and measuring the properties of the resulting interference pattern, it is possible to retrieve information about the night sky at spatial resolutions equal to the separation of the telescopes, overcoming the diffraction limit of a single telescope. In long baseline amplitude optical interferometry, the beams of light from the telescopes are transported to a central location and physically interfered. The interference is achieved via an instrument known as a beam combiner. In this review, I discuss the functionality of a beam combiner. I begin with a mathematical explanation of how interference fringes are produced and what information these interference fringes contain. This is followed by a discussion of how interference fringes are generated and measured in practise for the most common beam combination schemes, for both pupil plane and image plane combination and how these schemes can be realised in bulk optics or integrated optics. I also provide a detailed summary of the various design considerations that can affect the functionality of a beam combiner. Finally, I discuss current and future work in long baseline amplitude optical interferometry.</p>