At a meeting at work today a colleague mentioned that he is currently spending a LOT of time reviewing papers, it has reached a point where he felt that it is simply taking too much time from his other responsibilities. But on the other hand it is also part of the job so he did not really feel he could say no. I am (fortunately) not that much "in demand"; at the moment I am on average reviewing a 1-2 manuscripts a month and it is mostly short papers. I have also reached a point where I decline to be the referee if I think it will take me a long time to read up on the topic (meaning it is a topic I don't feel I know enough about) or if the paper is "too theoretical" (I am mainly an experimentalist) and requires a good working knowledge about some technique I am not familiar with. I've also become a bit better at recommending "do not publish" or "major revisions required" when the paper is simply too hard to understand because of the quality of the language (usually manuscripts from Asia); I have better things to do than deciphering text. In the past I guess I felt sorry for the authors and tried to point out at least the most obvious errors (which on a few occasions still ended up being a very long list); but that takes a very long time and I don't think that should be the job of the referee. The way I see it I should review the science not the text (not that my English is perfect; but if I am unsure about whether or not a sentence is correct I will ask someone who knows). It would be interesting to know how much time other people spend on reviewing papers? Also, what are your criteria for declining to be the referee?