What are the top 3 journals you would like to see break away from their commercial publisher?

I am looking for as many responses as possible. Please give the general field name (e.g. mathematics, geography, medicine). A brief explanation of the reasons for your choice would be nice.

Mine: (mathematics)

• Journal of Combinatorial Theory A (Elsevier) - I want to publish there
• Journal of Symbolic Computation (Elsevier) - ditto
• Inventiones Mathematicae (Springer) - it is probably the highest reputation math journal controlled by the commercial publishers