In short
When an accord fails, resist adding things. Describe the fault, rule out your own nose and your weighing, then use subtraction: remake it without one line at a time. Most accord problems have one of a small number of causes, and each can be tested with a small, cheap experiment.
What you will learn
- Describe a fault precisely enough to test it
- Use subtraction to locate a problem line
- Apply a sequence of checks before changing the formula
Before this lesson: Chypre Accord Workshop, Gourmand Accord Workshop, Musk Accord Workshop
Check these first
Three checks cost almost nothing, and they catch a surprising share of problems.
- Your nose. Come back tomorrow. Adaptation and fatigue make accords seem flat, and a tiring session makes everything worse.
- Your records. Check the arithmetic and the dilution of each line. A 1% stock used as if it were 10% is a tenfold error, and it would explain almost anything.
- Your materials. Smell each line fresh on a strip. An oxidized citrus or a contaminated vial can spoil a whole accord, and a cross-contamination problem will recur until you find it.
A diagnostic table
| Fault | Possible cause | Test |
|---|---|---|
| Muddy | Too many lines at similar levels; no leader | Subtraction. Then raise one line and lower another. |
| Flat | No contrast; low diffusion; fatigue | Rest. Then ladder in a small contrast. |
| Harsh or chemical | A strong line overdosed; a material off | Halve the suspect line. Check it alone. |
| Thin | Too little body; too much top | Compare with a larger share of the main material. |
| Falls apart over time | Weak handover between volatile and tenacious parts | Timed strips. Look for the gap. |
| Top dies at once | Very volatile top, or solvent masking | Check after the solvent has gone. Add a link material. |
| Changed overnight | A chemical reaction, or a contaminated vial | Remake fresh, and compare with the aged one. |
| Cloudy or separated | Insoluble waxes, a solid crystallizing, water | Check each line's solubility. Warm gently without flame. |
| Strong for you, invisible to others, or the reverse | Anosmia in one nose | Ask a third person. |
The subtraction test, worked
Suppose a five-line accord (A, B, C, D, E) smells muddy. Weigh five versions, each missing one line, plus the original. Label them without letting the label tell you which is which. Each has a small total, so this costs little.
Smell each against the original. Three kinds of result can appear:
- Removing a line improves it: that line is at fault, or overdosed. Try half the dose before deleting it.
- Removing a line collapses it: that line is carrying the structure. Protect it.
- Removing a line changes nothing: at this dose it does nothing you can perceive. Consider cutting it, since every line you keep has a cost.
If a formula has more than eight or so lines, group them, and remove one group at a time. Narrow it from there. This is the same logic as a binary search.
Know when a reaction is the cause
Some materials react. An aldehyde with methyl anthranilate forms a Schiff base over time, which changes the color and smell, and it is often done deliberately, as the Schiff bases guide explains. If a mixture has yellowed and shifted after a few days, make a fresh batch and compare. A difference between fresh and aged points to a reaction or to oxidation, and not to your composition.
Know when to stop
If you have made ten changes and it is getting worse, return to the last version you liked. An accord that cannot be rescued has still taught you something. Record it with a note on why you stopped. Knowing that a pair fights is a finding.
Practical exercise: Diagnose one failed accord
Choose an accord you were unhappy with, from any workshop in this level.
- Describe the fault in one specific sentence. Do not write "not good".
- Run the three first checks, and record the results.
- Use the table to choose a likely cause, and write it as a prediction: "If the cause is X, removing Y should improve it."
- Run a subtraction test.
- Make one revised version, then compare it with the original the next day.
You have done this well if your prediction was written before the test, and you recorded whether it was right. Being wrong and knowing it is a good result.
Common mistakes
- Adding a new material to cover a fault. It often hides the problem until later and adds another line to understand.
- Changing several things at once. You will not learn what fixed it.
- Trusting a judgment made late in a session. Check tomorrow.
Key terms
- Subtraction test
- Remaking a blend without one line at a time, to find which line causes a fault or carries an effect.
- Muddy
- A common bench term for a blend in which no part reads clearly, often from too many materials at similar levels.
Knowledge check
Written and reviewed by the RUŌOD Lab team. This is general education, not legal, regulatory, medical or safety advice. Example dilutions and exercises are for learning, are not validated products, and have not been safety-assessed. Completing a lesson confers no qualification. How we check these lessons.