This comment on Wikipedia suggests the exact composition may not be critical: " he compositions of the liquid in a storm glass varied but usually contained "camphor, nitrate of potassium and sal-ammoniac, dissolved by alcohol, with water and some air."
I suspect variation won't have much impact on it's spurious ability to forecast the weather; I suspect its appearance is a complex result of its recent 'thermal history'.
Neil
P.S. my Dad has one and considers it no more than a 'novelty'.