diff --git a/sp/seeing.png b/sp/seeing.png
new file mode 100644
index 0000000..1aa1780
Binary files /dev/null and b/sp/seeing.png differ
diff --git a/survey_performance.tex b/survey_performance.tex
index be8cc0b..8050a2e 100644
--- a/survey_performance.tex
+++ b/survey_performance.tex
@@ -19,11 +19,24 @@ \section{Survey Performance}
     \end{figure}
     
     
-Likewise, the survey simulations use a skybackground model as part of the model to determine five sigma visit depths and to choose observation pointings. The outputs available in the ConsDb include a `sky\_bg\_median` value, which is in counts per pixel. Together with an estimate of the platescale (0.2"/pixel) and a zeropoint, we can convert this into magnitudes per square arcsecond, to compare to the predicted values from the rubin\_scheduler skybrightness model. The results are shown in Figure~\ref{fig:sky}.  The values are also remarkably consistent, 
+Likewise, the survey simulations use a skybackground model as part of the model to determine five sigma visit depths and to choose observation pointings. The outputs available in the ConsDb include a `sky\_bg\_median` value, which is in counts per pixel. Together with an estimate of the platescale (0.2"/pixel) and a zeropoint, we can convert this into magnitudes per square arcsecond, to compare to the predicted values from the rubin\_scheduler skybrightness model. The results are shown in Figure~\ref{fig:sky}.  The values are also remarkably consistent, with a scatter of less than 0.15 magnitudes in all bands, and offsets within 0.1 magnitude except in y band, where the measured sky is 0.5 magnitudes brighter than expected. This is  within our expected errors in the skybackground model, particularly in y band where the sky is quite variable and harder to model. 
 
 \begin{figure}
     \centering
     \includegraphics[width=0.8\textwidth]{sp/sky.png}
     \caption{Predicted skybrightness values from `rubin\_sim.skybrightness`  compared to `sky\_bg\_median` converted to mags per sq arcsecond  from  from `cdb\_lsst.comcam.visits1\_quicklook`.}
     \label{fig:sky}
-    \end{figure}
\ No newline at end of file
+    \end{figure}
+    
+  
+We look forward to comparing seeing performance to survey predictions. Initial estimates indicate that the mean seeing for these visits was around 1.12 arcseconds, which isn't out of line with longer term survey expectations, especially given that we remain in the early stages of commissioning. 
+
+\begin{figure}
+    \centering
+    \includegraphics[width=0.8\textwidth]{sp/seeing.png}
+    \caption{Predicted skybrightness values from `rubin\_sim.skybrightness`  compared to `sky\_bg\_median` converted to mags per sq arcsecond  from  from `cdb\_lsst.comcam.visits1\_quicklook`.}
+    \label{fig:seeing}
+    \end{figure}
+    
+   
+Remaining questions include the efficiency of observations, and in particular the likelihood of whether a single snap will be sufficient. 
\ No newline at end of file