Sorry for the delay in responding, we are now figuring out why we didn't get a notification that this was posted.
However, to answer your question, what do you believe is incorrect about the map? Which settings did you use on the kernel density analysis? The key issue here may be the size of the area that you're looking at relative to the size of the cells. When looking at large areas with a single point of high concentration relative to the entire area, you're likely to run into this kind of clustering.
There are two suggestions I'd make right away: increase the # of cells you are using (or decrease the unnecessary area around the cluster if you want more definition) and then consider using a kernel function with a larger drop-off rate. If you hold the option key down when choosing the analysis, you will get access to a large number of additional parameters that allow you to change how the analysis is performed.
Cheers,
-Gaige