The best reference on citation of email messages and articles is Section XII of Kathy Patrias' "National Library of Medicine Recommended Formats for Bibliographic Citation" (NLM, Bethesda, MD, 1991, PB 91-182030/GBB, $28.50; (703) 487-4650). "Include Author (affiliation, optional), Title, Connective Phrase (Message to:) Recipient, Connective Phrase (In:) Title of bulletin board or mail system, Type of medium, Author of mail system (omit if same as publisher), place of publication, publisher, date of publication of message, numeration of message, pagination of message, availability, and language." Example: "Woodsmall, Rose. Charging for searches. Message to: John Faughan. In: GRATEFUL MED BBS[electronic mail system]. [Bethesda (MD)]: National Library of Medicine. 1990 Nov 2, 10:05 am; Message No.: 11356; [6 lines]. Available from 800-525-5756." [Gil Hopson ([email protected]), PACS-L, 6/29/93.]

Other style manuals (APA, Chicago, ALA) say that you shouldn't cite things that are not retrievable. Citing a person or personal correspondence means obtaining permission to go "on the record" and asking what contact information they're willing to make public. [Greg Newby ([email protected]), VPIEJ-L, 2/16/92.] (I try not to offend anyone, but I find this too strict. Posts to the net are often saved and quoted, even when there is no official archive. The number of "witnesses" make newsgroup postings analogous to public speeches. A reporter should have rights to fair use without asking permission, but must obviously cite the source.)