deletions | additions       diff --git a/As_can_be_seen_in__.tex b/As_can_be_seen_in__.tex  index 84ffe00..ed789d4 100644  --- a/As_can_be_seen_in__.tex  +++ b/As_can_be_seen_in__.tex 
...
As can be seen in Figure~\ref{fig:fermidirac} the probability falls to 50\% of the maximum at $\mu$. The width of the attenuation of the probability is defined by the $\sigma$ value, such that smaller $\sigma$ values mean that the probability falls off more quickly (still centred on $\mu$). If we say that $r=\mu/\sigma$, and decide to define the distribution in terms of $\mu$ and $r$, then we can empirically estimate the attenuation band as a function of these parameters. The band over which the probability falls from 97.5\% of the maximum value down to 2.5\% is given by $\mu \pm Z\mu^2/2r$, Z\mu/2r$,  where $Z \approx 0.349$. 7.33$.  So, for example, if have $\mu = 10$ $\mu$  and we want the distribution to fall from 97.5\% of it maximum down to 2.5\% over a range of $\mu \pm 0.1\mu$ X\mu$  then $r \approx 1.74\mu$ Z/2X$. Or, plugging in the numbers for $\mu = 10$ and $X = 0.1$, gives $r \approx 36.7$,  or $\sigma = 0.57$. \mu/r = 0.27$.