Fourier spectral analysis for unevenly spaced, average value, data

We propose a method for Fourier spectral analysis of unevenly spaced data which correspond to averages of a given quantity; the method is particularly adapted to data from counting, such as specimen counts from borehole samples. The method is extremely straightforward and takes into account the effects of data averaging. A Monte Carlo approach is used to evaluate possible effects of random errors in time determinations and in counting; probabilities for different possible spectral values are estimated. The method is implemented as a Matlab function.