Reference
Reference: records that make an experiment reproducible
14 November 2026 · 5 min read
A result becomes evidence when somebody else can say exactly what was in the tube. Every record below exists for that one purpose: to keep an unbroken line from a number in a notebook back to a defined quantity of a defined material. None of it is interesting on the day it is written, which is why it is so often skipped.
The chain, link by link
Reproducibility is a chain, and each link has to hold:
- the measurement, as recorded
- the aliquot the measurement was made from
- the solution that aliquot was drawn from
- the vial that solution was made from
- the lot that vial belongs to
- the analytical documentation issued for that lot
Break one link and the rest carries no weight. The observation still happened, but it can no longer be attributed to a defined material, and an observation that cannot be attributed is an anecdote.
Why one missing field is enough
The record set earns its place on the day two results disagree, which is usually months after both were produced.
With complete records you can ask testable questions. Did the two runs use the same lot? Was the solvent the same, down to its lot? Had the aliquot been thawed once or five times? Each question has an answer in the log, and each answer either explains the difference or eliminates a cause.
With one field missing, the only explanation available is that something changed, which cannot be tested, and the earlier work becomes a story rather than data.
The minimum record set
| Field | What to write down | Why it is there |
|---|---|---|
| Material | full name and catalogue reference | an abbreviation that is obvious today is a riddle in April |
| Lot number | copied character for character from the vial | the only link to the documentation issued for that batch |
| Date received | the day the parcel arrived | fixes the start of your custody of the material |
| Condition on arrival | packaging intact, vial intact, cake present | separates a supply problem from a bench problem later |
| Storage since arrival | where it went and at what temperature, with dates | a gap here is a period nobody can account for |
| Date of reconstitution | the day the solvent went in | the clock on a solution starts here, not at purchase |
| Solvent | type and its own lot number | a change of solvent is a change of experiment |
| Volume added | the volume actually delivered | the recorded concentration depends on it |
| Concentration | the figure derived from mass and volume | saves reconstructing the arithmetic from memory |
| Aliquot scheme | number of portions, volume of each | tells you what a tube holds before you open it |
| Storage location | freezer, box and position | a precise record beats a search through four boxes |
| Thaw count | one mark per entry into the tube | the history that explains a weak result |
| Operator | initials against every entry | tells the next person who to ask |
| Deviations | anything that did not go to plan | the field most often blank and most often needed |
The solvent is a material in its own right
A solvent has a lot number, a date of opening and a storage condition exactly as the peptide does, and it is the variable people forget because it feels like a background ingredient. Bacteriostatic water contains a preservative, sterile water does not, and a dilute acid solvent changes the pH. Two runs that used different solvents are not replicates of each other.
Write the solvent lot on the same line as the peptide lot, with the date the bottle was opened.
Thaw count and the aliquot scheme
Record the scheme at the moment the portions are made, not afterwards: how many tubes, what volume in each, where they went. Then add one mark to the tube and to the log every time it is entered.
A tube thawed an unknown number of times is a tube of unknown history. It can still be used, but any result from it carries a caveat that cannot be resolved.
Written at the bench, or written as fiction
Records written an hour later are written from memory, and memory supplies the volume you intended rather than the volume you delivered. If the pipette stopped at 4.8 ml, the log says 4.8 ml, and the concentration is recalculated to match. A concentration calculator removes the arithmetic as an excuse for a blank field.
If the records must end up in a spreadsheet, write at the bench and transcribe afterwards, keeping the original.
What goes wrong
- The lot copied “close enough”. One transposed character and the vial no longer matches its analytical documentation.
- The plan recorded instead of the event. A log of intentions is a log of nothing.
- “Water” as a solvent entry. Three different materials answer to that word.
- A storage gap. Material that sat somewhere unrecorded between the courier and the freezer.
A short checklist
- One lot number, copied character for character, on arrival.
- Storage recorded from the date of receipt, with no unaccounted gaps.
- Solvent type and solvent lot on the same line as the peptide.
- Volume delivered, not volume planned, with the concentration derived from it.
- A thaw mark per entry, and initials against every line.