The context of this study is related to the cyclus of gas and dust in our 
galaxy. What is the composition of stardust produced by various types of 
stars in our galaxy? How does it enter the ISM and how is it modified, and 
what is the composition of this material at the onset of star- and planet 
formation? Crystalline silicates are a useful tracer of the way dust gets 
cycled in the ISM, because it has strong and well defined spectral 
resonances in the mid-IR. We know from ISO and Spitzer observations that 
evolved stars produce crystalline silicates, but strangely enough we do 
not find it in the diffuse ISM. It is present again in the immediate 
surroundings of young stars. What happens in the ISM? How much crystalline 
stardust is really produced? New Spitzer observations of evolved stars in 
the galactic bulge allow a detailed inventory of the dust production of AGB 
stars. The most important contributors to oxygen-rich dust are OH/IR stars 
because they have the highest mass loss rates.