Kernel density distribution where range is polar

Hi there,

I am not sure if the question heading makes sense but I will try to explain my problem more clearly here. I have a set of data d and I am calculating the density estimation over the range x using the following code

x=linspace(0,0.259*pi,201);

[f,xi]=ksdensity(d,x);

However, the range corresponds to the circumference of a circle and clearly this creates a discontinuity. Does anyone know of a way of avoiding this discontinuity so that the density estimation at each end of the range is updated if an event occurs near the discontinuity? I hope this makes sense.

Many thanks in advance

Jack