Finding Extreme Subdwarfs

I develop a new technique to identify M-type extreme subdwarfs (esdMs) and demonstrate that it is substantially more efficient than previous methods. I begin by obtaining spectroscopy and improved photometry of a sample of 54 late-type halo candidates using the rNLTT reduced proper motion (RPM) diagram. From spectroscopy, I find that four of these are esdMs, three of which were previously unknown. From the improved photometry, I show that all four lie in a narrow RPM corridor that contains only 4 non-esdMs. Hence, with good photometry (i.e., without spectroscopy), it appears possible to select esdM candidates with a 50% esdM yield. This is more than an order of magnitude more efficient than previous methods.