Cataloguing a pair of agate halves as one specimen

You buy a pair of agate halves. The dealer has them set out on the table, two faces polished, the banding running across both halves as a clean mirror. He's priced the pair as one specimen. You pay for it and he wraps the halves together.

So how do you record it?

If you make one record, the dimensions field can hold one length and one width and one height. You could measure either half, but the two halves aren't identical — one is a touch thicker than the other, and the polished face of the larger half is two centimetres wider than the smaller one. None of the numbers you type into one row are the truth for both pieces.

If you make two records, you get two collection numbers for what the dealer sold you as one specimen. When the insurance schedule needs to add up, you're counting two things that shouldn't be two things. When someone reads your catalogue in twenty years, your records say MIN-0042 and MIN-0043 with nothing in between to say they belong together — and the original pair note, if it survived at all, is buried in a free-text field on one of the two rows.

Neither answer is right.

The trap the spreadsheet sets

Most collectors hit this for the first time when they're a few hundred specimens in. They've been using a spreadsheet because a spreadsheet is what every collector starts with. One row per specimen, columns for the fields. The columns are length, width, height, weight, locality, acquisition date, cost.

The spreadsheet has no answer for "two sets of dimensions on one row." So the collector improvises. They put the larger half's measurements in the main columns and stick the smaller one in a notes field, like "second half 92×68mm, 142g." It works at the moment of typing, but six months later they filter the catalogue for specimens under 100mm wide and the smaller half doesn't show up because its dimensions are buried in a notes string.

The same collector has tried the other workaround. Two rows, each with its own collection number, each with its own dimensions. Search works. The insurance schedule adds up correctly if you remember to mark both rows as sold when the pair sells together. They add a "notes" column to flag that MIN-0043 is the matched half of MIN-0042. Six months later they've made a "set ID" column to make those flags searchable. A year in they've reinvented half a relational database in three columns, and they're the only person who can read it.

The problem is that the spreadsheet has one row per thing but the dealer sold one specimen that happens to be two things. The model is wrong. Adding columns won't fix it.

How experienced collectors handle it

The pattern that works is to treat the pair as one specimen with two pieces attached to it. One collection number for what the dealer sold you. One row in the catalogue. But the row carries a list of pieces, and each piece records its own physical detail.

Concretely:

• One specimen record. One mineral name (Agate), one locality, one acquisition date, one cost. One collection number, like MIN-0042.
• Inside that record, two pieces. Each piece has its own name ("Left half" and "Right half" works, or "Outer" and "Inner" if you prefer), its own length, width, and height, its own weight.
• The top-level dimensions and weight on the parent record describe the aggregate — combined weight of both halves, the bounding box of the assembled pair. They're meaningful, just at a different scope.
• The label that hangs from the cabinet carries the parent record's identifying fields. One label per specimen, not one per piece.

The mental model: the specimen is what the dealer sold and what the insurance schedule values. The pieces are the physical objects inside. Both are real, and the catalogue records both at the level each actually describes.

This isn't only for agate halves. The same shape fits a matched pair of geodes, a thumbnail with its detached matrix, or a polished mirror pair from a stone slab. The trigger is "one acquisition, two or more physical pieces that travel together as one specimen."

A few things that catch collectors out when they first try this model:

Decide early what changes when the pair is split. Most pairs stay together for life, but sometimes a half sells alone. If that happens, the specimen needs to either drop that piece (the catalogue now records one piece, and the missing one shows up in the sale record) or get split into two specimens with new numbers. Both are valid; the catalogue should record what actually happened, not pretend the specimen was always two separate items.

Photograph the pair together. When you photograph this specimen, at least one shot needs both halves visible side by side. That single image is what a future cataloguer (or you, three years from now) needs to read "one specimen, two halves" at a glance. The individual portraits of each half are useful for detail; the pair shot is the one that defines the specimen.

What this looks like in ClearBench

This is the model ClearBench's Pieces tab is built around — purpose-built mineral collection software for cataloguing specimens that arrive in more than one piece. On the item view, the Pieces tab carries a separate row for each piece with its own dimensions, weight, colour, and condition fields. The schema editor lets you rename the noun ("Halves" if "Pieces" doesn't fit your taxonomy) and customise the field set per type, but the default field set covers the size-and-weight case directly.

The collection number, mineral name, locality, and label still live at the specimen level. The pieces are inside the specimen, not next to it. One specimen renders one label. The insurance schedule values the specimen as one item, not two. Search across the catalogue still finds the specimen, and a future feature can also surface the per-piece dimensions when you want to filter by "any piece under 100mm."

If you'd rather not use ClearBench, the same model works in a more disciplined spreadsheet: one sheet for specimens, a second sheet for pieces with a foreign-key column pointing back to the parent specimen, a third sheet for sales. Most collectors who hit this problem and don't have a tool to model it end up building exactly that, and they stay on that pattern for years.

The tool matters less than the model. Once you stop trying to fold a multi-piece specimen into one row, the catalogue stops fighting you.