How does the gap change at $T_c$ in underdoped cuprates?

Many measurements on underdoped cuprates have shown a gap that persists up to room temperature. This raises an important question: what happens at $T_c$ in order to cause the loss of perfect conductivity? In ARPES, the nature of the gap changes from d-wave below $T_c$ to Fermi arcs above $T_c$. However, ARPES necessarily averages over significant nanoscale disorder. We will present detailed STM spectroscopy on underdoped $\textrm{Bi}_2\textrm{Sr}_2\textrm{CaCu}_2\textrm{O}_{8+x}$ from both single points and areal averages. By using a local probe we avoid averaging over the disorder. We have performed lattice tracking spectroscopy on identical atomic sites [1] and indentical grids of points [2] for a range of temperatures both below and above $T_c$. Unlike overdoped samples, the STM spectrum in underdoped cuprates shows two energy scales [1]. We will compare our data to models based on ARPES, with emphasis on the difference between the superconducting and pseudogap phases. [1] Gomes \textit{et al., Nature} \textbf{447}, 569 (2007) [2] Pasupathy \textit{et al., Science} \textbf{320}, 196 (2008)