Green OA and the role of repositories remain controversial. This is perhaps less the case for institutional repositories, than for subject repositories, especially PubMed Central. The lack of its own independent sustainable business model means Green OA depends on its not undermining that of (subscription) journals. The evidence remains mixed: the PEER project found that availability of articles on the PEER open repository did not negatively impact downloads from the publishers’s site, but this was contrary to the experience of publishers with more substantial fractions of their journals’ content available on the longer-established and better-known arXiv and PubMed Central repositories. The PEER usage data study also provided further confirmation of the long usage half-life of journal articles and its substantial variation between fields (suggesting the importance of longer embargo periods than 6–12 months, especially for those fields with longer usage half-lives). Green proponents for their part point to the continuing profitability of STM publishing, the lack of closures of existing journals and the absence of a decline in the rate of launch of new journals since repositories came online as evidence of a lack of impact to date, and hence as evidence of low risk of impact going forward. Many publishers’ business instincts tell them otherwise; they have little choice about needing to accept submissions from large funders such as NIH, but there has been some tightening of publishers’ Green policies (page 102).