The first sections of this article are well written and accessible to a wide readership. However, the presentation overall could still be improved. For instance:
1. Historical introduction. Too much emphasis on "ice ages" modeling (the so-called 100kyr problem in Milankovitch theory). At this stage the role of SR in climatology is not well established, yet, if not disputed altogether. On the contrary, the tremendous success of SR in the current literature is mostly due to its applications to biology and neurobiology, e.g., by the St. Louis group. I think this fact should be clearly stated in the introduction (and not incidentally, at the very end).
2. Residence time distribution. This characterization of SR as a synchronization mechanism had a great impact in the biology literature. I think that the original paper by the Perugia group should be added to the References (Gammaitoni et al, PRL 62 (1989) 349.
3. Stochastic resonance beyond the classical setting. This part is too sketchy to be of any practical use, unless links are given to the relevant Wiki- or Scholar-pedia pages. I understand that those are technical issues, but their presentation, if deemed necessary, should read more plausible to a non-expert reader (say a student). If too complicated, better drop a few item from the list.
I think this is a well-written and useful article. I do agree with the other reviewer comments above this section, although I don't think the discussion of the role of SR in climate is overly long. This was the context of the introduction of the concept, after all. I also agree wholeheartedly that the role of SR in synchronization should be briefly discussed. In addition to these comments there are a couple of other things I would like to see done before I approve the article ( I have correcetd a few typos and added a few commas as well):
(1) I would like to see more specific referencing to the articles cited. The Scientific American/trade style of writing is nice, but this is a more technical encyclopedia and most other articles cite their references more carefully in the text (this includes webpages where relevant). This allows interested readers to more easily pursue more detail on the topics.
(2) I would like to see some material added on Fisher information and mutual information in the "Further indicators..." section. Both of these measures have been used extensively in the more recent literature and are arguably more useful for the more exotic forms of SR (see, e.g., Greenwood, Ward, Wefelmeyer, 1999, Phys. Rev.E).
(3) In the section "...beyond the classical setting" I would like to see mention of nonperiodic (Collins & Chow), threshold (non-dynamical) (Gingl, Kiss & Moss), and suprathreshold SR (Stocks). The authors' emphasis on dynamical SR is understandable, and may even represent a position that non-dynamic SR does not deserve the name. In the last 10 years, however, most of the work in biological, neuroscience, and psychological applications has been in the context of these concepts. A link could be added to the Wilkens & Moss Scholarpedia article on Mechanoreceptors and threshold SR.
(4) A useful reference summarizing a lot of the more recent work on SR in biology, perception, and neurophysiology is the tutorial article by Moss, Ward & Sannita (2004), Clinical Neurophysiology, 115, 267-281. This could be referenced in the "Experimental aspects...." section.