Robert Pries wrote, > >most of us want some information about a particular bulb species. ... at > the beginning of each article a list of keywords. If one placed these > three or four words in searchable fields, then one could still present > the author title information but have a way of searching the data for > special topics. This works for short papers. But what about Floras and > the like. You would have to have a field for every species listed to be > of value. I was not proposing a simple list; I mentioned that the bibliography would be "indexed" at least by genera. The index could simply be an editable text document, separate from the list but referring to its entries either by unique numbers or by some other convention (such as author/date); or it could be integrated with the list in a single database. This is what we need to decide, and what I need help in designing. It would be most useful to list every species in the index. The hardest reference task for me is looking up obscure names, and these often turn out to be invalid names. In the sample index entries I've been playing with, I decided to list synonyms in the format "[Cyclamen] neapolitanum = hederifolium"; the equal sign is more efficient than "see", the usual convention for index cross-reference. I haven't been putting subspecies and varieties in the index at this point, on the assumption that this tracking these down is work that can be done by the user. I've proposed annotations (very brief, objective ones) be included in the bibliography entries, so the user can decide, for example, that she is more likely to find Crocus biflorus ssp. crewei in Mathew's monograph of that genus than in his "Complete Guide to Growing Bulbs," even though C. biflorus is indexed to both. Furthermore, mentioning that Phillips & Rix illustrates all the indexed plants will tell the user that a photo of the subject can be found there, though perhaps not anywhere else. Some indexes use a typographic convention to signal illustration, but this is probably more work than we want to do. Regarding Jim Shields's proposal for a relational database: Aren't such databases extremely space-consuming, with a "different table for each taxon" as Jim writes? Hope this clarifies my thinking. Jane McGary