Stephen,

I've always kept a hand-written catalog of my shells. It's written like
a table with following columns:

Catalog Number (starting with 1 and working up incrementally...
            no special system is easiest and most versatile)
Species Name (genus, species, author, date...written in pencil
            because this can change over time)
Locality (general locality data as to the place it was found...not the
             habitat or conditions...that goes in the Habitat/Remarks column)
Date Catalogued (just for my own knowledge of when I acquired it)
Number of Specimens (all specimens collected at the same time in the
             same location/habitat are put together into the same lot or
             catalogue number)
Date Collected
Collector
Source (who donated it or who I bought if from...same as Collector if I
             collected it myself)
Size
Grade (G, F+, F, etc)
Habitat/Remarks (the most important column...along with Locality. This
             contains any remarks about the conditions the shell was found
             in, or anything else that I want recorded. eg. "Found at lowest
             tide level in mussel bed with Mytilus edulis and Balanus glandula,
             on fully-exposed rocky headland. In crevies and under rocks
             out of sun on hot sunny day..." "with operculum" "preserved in
              alcohol", etc)

The shells themselves, if large enough, get numbered by applying clear
nail polish, then the number with Indian in (fine calligraphy pen), then
another
layer of clear nail polish. This way the shell is permanently numbered, but
the number is easily removed with nail polish remover, leaving no mark.

Ideally, each lot of shells should have a data slip in the tray with it so that
you may easily refer to it, but all you really need it the catalog number with
the shells, so you can look it up in the catalog. I've been meaning to
additionally
create a database for cataloging my shells, so that I can easily access data
like how many I have of a species or family and make reports as such, but at
this time all I have is a written book that I record such things in: a
page, or more,
for each family, with the species names and the catalog numbers and total
number of specimens for each species recorded. The catalog, though is the
most valuable item for my collection and even if I had a computer database,
I would still keep the written catalogue for security purposes...what if
your computer
crashes and you loose the entire database!!!!

As far as number of specimens per species: I like to get as many specimens
of each species that I can get my hands on, although, when I'm collecting
I would never think of taking too many from one habitat at one time! A species
that shows little variation can be represented in your collection my just a few
specimens, but for others that show great variation, your collection would
benefit from having as much variety as possible, especially with good locality
data, so that one may study the species variety as it pertains to locality and
population. I have many hundreds of Nucellas (4 species), because they are
a varied and confusing bunch, but only 3 Strombus gigas, because they are
big and take up too much room to store, and show very little variation.

I hope this helps,

Peter Egerton