Estimating missing points using Fourier series

 I have a set of (seasonal evaporation)data with some missing points. How can I estimate the missing values by using Fourier interpolation (or approximation)? 
 The function "Interpft" (Fourier transform interpolation) in the toolbox only works with "equally spaced" points (as it's mentioned in MATLAB help), and as I said I have some missing points so they are not "equally spaced".
 For example I have: 
[2.27  6.17  8.44  4.57  3.01  6.22  9.24  4.78  ----  6.11  7.98  ----  ----  5.34  8.28  4.59] 
for winter 2002, spring 2002, summer, autumn, ...., summer 2005, autumn 2005 respectively, and I want to estimate those 3 missing points (season 9, 12 and 13) by using Fourier series.