I've been enjoying the discussion about High Energy Physics (HEP) publishing and the arXiv, and wanted to make a few observations.
1) One piece of information in the recent work of Gentil-Beccot, Brooks and Mele http://arxiv.org/abs/0906.5418v1 is that papers put on the arXiv and never published in a peer reviewed journal are cited 1/6 as much as those that are eventually published. In addition, papers submitted to peer review journals that were never put on the arXiv also suffer the same low citations.(see Figure 2). The American Physical Society, publisher of Physical Review D, has embraced and supported the arXiv, as it provides the "newspaper" for the field, while the journals provide the validation. Both are necessary. 2) Our experience at APS is that we have not seen significant loss of subscriptions of PRD as a result of the arXiv. PRD is the largest of the HEP journals, with currently more than 45% of the HEP literature. 3) The real question is "Why have we not lost subscriptions to a journal that is 97% on the arXiv". Note that this high fraction on the arXiv is not true for the rest of physics. Our other journals have typically 40% of the articles on the arXiv, and the AIP Applied Physics journals are much less. ( A detailed study was recently made by TIm Ingoldsby:http://www.councilscienceeditors.org/events/annualmeeting09/presen tations/Ingoldsby.pdf ). There are several reasons we can think of for the lack of cancellations, such as: Our version of the paper is definitive. Librarians subscribe to have guaranteed access to the 3% which is not on the arXiv. Institutions support the peer review process, by supporting the journals. Faculty are loyal to APS and oppose cancellation. Librarians appreciate that we are committed to preserving our content archive, and support us. However, I don't think any of these is as important as are the following two points: A) We provide a package of all of our journals, APS-ALL. The cost of the package, and the cost of buying separately all of the remaining journals without PRD are not substantially different. I guess we could call it a "little deal". The APS journals form the core of any physics collection. B) PRD is the largest journal in HEP and is also the cheapest/ article. Our price/article and those of the other society publisher, SISSA, are significantly smaller than the other journals in the field. The price/article listed in journalprices.com are: PRD $1.88, JHEP 2.19, PLB 11.79, EPJC 24.22, NPB 36.63. (Because of "deals" these prices are not necessarily what is paid for them, but they do reflect the scale). Although the fundamental economic problem is that an author's choice of journal does not strongly depend on the price of the journal, HEP authors are responding slowly, and the fraction of HEP literature in PRD and JHEP has grown from 35% in 1997 to 66% in 2007. Cancellation information in other journals is not available, but the fact that PRD has not been cancelled as a result of the arXiv may be because of our package and our low cost, and should not be generalized to all HEP journals, and to the effects of self archiving in general. 4) APS is committed to making our journal content as widely accessible as possible, consistent with sustaining ourselves so that we can continue. Our model has been to be as efficient as possible so that we can have our content available cheaply and therefore broadly. Indeed, a significant fraction of our subscription revenue comes from smaller institutions that are not generating significant numbers of articles. If we were to convert to an author pays open access model,then the larger institutions that generate the research articles would have to pay much more than they are paying now, at least for APS journals. Our experience with author pays "page charges" in the 90's was disastrous, and it almost killed PRD, and required us to eliminate page charges for all of our journals, except PRL. This is unfortunate, as it could have naturally led to open access without subscriptions if the community had supported it. The SCOAP3 consortium offers an opportunity to make the transition to author pays, but the question of sustainability of the funding model is not yet settled. Sincerely, Gene D. Sprouse Editor in Chief, American Physical Society