He got his career started, strangely enough, by studying motion sickness in dogs. His most famous discovery, though (motion sick dogs notwithstanding) is the columnar organization of cortex. Obviously he did much more. Here is a chapter on The History of Neuroscience devoted to Mountcastle which is well worth reading. You can also watch the video (embedded below). He evidently was also the first president of the Society for Neuroscience (SfN), a position he ran for on a lark.
Here is an interesting bit about the discovery of the columnar organization of cortex, and its reception:
Many friends have inquired why the description of this general principle is contained in the paper authored by me alone. The answer is: by request! My two colleagues were so apprehensive over my proposal of such a radical hypothesis that they sought to disavow themselves from it! Indeed, it is not possible to exaggerate the calumny I was subjected to over this proposition, and with the most vigor by my colleague Jerzy Rose. He and most other anatomists had been trained in the schools of Nissl cytoarchitecture, Rose by the Vogts themselves, and the idea of layered cytoarchitecture dominated the scene; some even designated different layers for different functions! All this was before the revival of Cajal-type studies of the cortex. One critic said that the idea was just the “musings of an old man,” and I was only 39! Columnar organization was confirmed in a few years for the visual cortex by Hubel and Wiesel, and then by many others for the homotypical cortex as well, and it is now part of the cortical zeitgeist.
His most-cited papers are:
Modality and topographic properties of single neurons of cat’s somatic sensory cortex J. Neurophysiology 1957
Parietal lobe mechanisms for directed visual attention J. Neurophysiology 1977