deletions | additions       diff --git a/subsubsection_Choice_of_row_separation__.tex b/subsubsection_Choice_of_row_separation__.tex  index a7fd1da..1fd0056 100644  --- a/subsubsection_Choice_of_row_separation__.tex  +++ b/subsubsection_Choice_of_row_separation__.tex 
...
\subsubsection{Choice of row separation}  The nominal separation given in the TIP of $\theta_{row}=7.25^\prime$ was derived by setting a uniformity specification of 1\% or better at 3.6GHz assuming $\theta_P=12.5^\prime$. However, this value does not produce an integer number of rows over a convenient span (say integer number of degrees). It would be beneficial for the tiling if it were to do so. In addition, recent beam mapping observations have measured the S-band beam to be narrower than previously assumed (\S~\ref{sec:setup-surveyspeed}), indicating that correspondingly narrower row separations are needed to limit the error to the same extent as discussed above in \S~\ref{sec:setup-scanrate}. Obvious options are to use $6.67^\prime$, $7.2^\prime$, or $7.5^\prime$ instead. Implications of these changes on the scan speed and mosaic errors using the values in Table~\ref[tab:otfuniform} Table~\ref{tab:otfuniform}  and discussion in \S~\ref{sec:setup-scanrate}) are: \begin{itemize}  \item $\theta_{row}=6.67^\prime$ : There are 9 rows in $1^\circ$ for a total of 810 in the North, and an  additional 360 rows in the South to get to $\delta = -40^\circ$, plus one at the equator, giving a total of 1171 rows in the survey. The scan rate would be $\dot{\theta}=3.573^\prime$/sec for a survey speed of $SS=23.83$~deg$^2$/hour.