The short answer is because of the changes in reading nature of scholars switching to online search.
Most researchers find their sources via online search, which is facilitated by metadata descriptions that accompany journal articles and increasingly books. However, there is one category of scholarly work that is gradually becoming obscure due to lack of metadata of its own, that is chapters of an edited compilation. Of course as a book a collection of essays would have its own metadata, that acknowledges the editor and provide a general description of the book. However, the individual chapters unlike journal articles are not provided with independent descriptions of their content via their own metadata that would allow them to be found. They are locked inside a book that gives them poor visibility to their potential readers.
In addition as a publisher we are finding it increasingly difficult to peer review edited compilations to the same standard that we do so for journal articles and monographs. They come as part of a package that the book editor has curated to convey a theme. However, what do we do when some chapters do not meet the peer review standard while others do. Removing the lower quality chapters destroy the theme for the book. Should we reject the whole book just for a couple of poorer chapters?
As a responsible academic publisher we have been trying to find solutions to these issues. Our first reaction was to reduce the number of edited compilations we accept. However, this has not been popular with our community who still see this as a good way of reaching an audience quickly. In particular editors of such books do not lose anything, it
is the contributing authors who are short changed.
An alternative is to persuade editors to guest edit a journal on the proposed topic of their book first before the material is reissued as a book in its own right. In this way we address both problems of online access and peer review in one go. The articles being part of a journal get an independent DOI (Digital Object Identifier) and metadata so they can be individually accounted for and found. Also the journal publication will add the peer review process as a matter of course and the book editor will be left with high standard of articles to curate them into a book later.
While some of our community have accepted the merit of the above solution others find the delay in the publication cycle (that pre publication in a journal brings with it) unacceptable. Most of our journals have material in their pipeline and would fit a special issue within a 2 year time cycle. Our book programme on the other hand has a 12 months turn around from a successful peer review.
As a third option we have been looking at is the possibility of publishing edited compilations as if they were journals, giving their chapters metadata and DOI via hosting them on our journal platform. This may increase costs, not solve the peer review problem but offer a solution satisfactory to some. However we have seen some editors object to this as they see their anthologies as multi authored books that should be read from beginning via the middle to the end. These editors see disaggregation of the book as detrimental to the coherence of their efforts in putting the collection together.
While as publishers we may not be able to satisfy the needs of every scholar in the same way we shall try and address the challenges of change in technology and reading habits.