Blending
Blending by weight on a scale that reads to a hundredth
The first blend I liked, I could not make again. I had put it together with a teaspoon and a vague sense of proportion in the autumn of 2022, and the four attempts that followed were all recognisably related to it and none of them were it.
Why a spoon cannot hold a blend together
A level teaspoon of a fine broken black leaf weighs a little over two and a half grams on my scale. A level teaspoon of a large open-leaf green, the sort that looks like dried hedge clippings, weighs about nine tenths of a gram. A teaspoon of dried flower petals weighs less than half of that, and is so compressible that how hard I pressed the spoon into the jar changes the figure by a third.
So a blend written as two spoons of one thing to one spoon of another is not a ratio at all. It is a ratio for one particular pair of materials on one particular day, and as soon as the leaf settles in the jar or I buy the next batch with a slightly different cut, the same spoons give a different blend. That is why my first good blend was irreproducible: I had never recorded what it was, only what I did.
Weight fixes this completely and immediately, and it is the single change that made blending an activity I could learn from rather than a series of lucky accidents. It costs nothing except the discipline of putting a bowl on a scale and pressing tare.
The scale, and what it does when you are not watching
Mine is a small flat-bed scale that reads to a hundredth of a gram up to two hundred grams, of the kind sold for weighing coffee and powders. It was not expensive and it is not laboratory equipment, and it has three habits worth knowing about.
It drifts when cold. From switch-on, the reading creeps by two or three hundredths over the first four or five minutes, so I turn it on before I fetch the jars and I re-tare immediately before the first component goes in. It also drifts slowly across a long session as it warms, which is why I tare between every single component rather than adding cumulatively and trusting the running total. And it is honest only within its own resolution: anything under about a tenth of a gram is a number it will print but not a number I would stake a blend on. Below that I scale the whole batch up instead.
The surface matters more than I expected. The table by the window is on old floorboards and somebody walking across the room shifts the reading by a hundredth or two. I moved the scale to a flat stone chopping board and the readings steadied noticeably. A draught from an open window will do the same thing, which I discovered while blaming the scale.
Kept on the card
- Switch on, wait five minutes, then tare; a cold scale reads low
- Tare between every component rather than trusting a running total
- If a component works out under a tenth of a gram, make a bigger batch
- Put the scale on something heavy and flat, and close the window
Sixty grams, written as percentages
I build in sixty gram batches. That is large enough for the small components to land on numbers the scale can hold, around twenty cups of drinking, and small enough that a failure costs me a fortnight rather than a season. The recipe goes in the notebook as percentages, never as grams, so that scaling it up or down later is arithmetic rather than a new experiment.
A blend I have made eleven times now runs at seventy-two percent of a medium black leaf, twenty percent of a lighter and more floral black, six percent dried citrus peel and two percent dried cornflower, which is there entirely because it looks well in the jar and contributes nothing I can taste. In grams at sixty that is 43.2, 12.0, 3.6 and 1.2, and every one of those is comfortably within what the scale can do.
The components go into a wide bowl in descending order of quantity, tared between each, then get turned over with the hands rather than stirred. Stirring with a spoon breaks the larger leaf and pushes the fines to the bottom, and a blend that has separated by particle size will give a different cup from the top of the jar than from the bottom. Turning it by hand for about a minute, lifting and folding, keeps the cut intact.
Two weeks in the jar before judging it
A blend tasted the hour it is made tastes of its loudest component. Anything aromatic sits on the surface of the other leaf and has not yet distributed itself, so the first cup from a fresh batch is not representative and, in my experience, is misleading in a specific direction: it always makes me think I have added too much of whatever smells strongest.
I now leave a new batch sealed for two weeks before deciding anything, in a full jar with as little air above the leaf as I can manage, and I taste at day one, day seven and day fourteen, writing all three down. The day one and day fourteen notes for the same batch are often so different that if I did not know it was the same jar I would not connect them. Where a citrus or a spice is involved, day fourteen is consistently the gentler and better cup.
The cost of this is patience and the benefit is that I stopped correcting blends in the wrong direction. Three of my early attempts were adjusted downwards on the aromatic component after a day one tasting, and all three ended up thin and characterless a month later.
The batch I ruined, and what it taught me about small numbers
In January 2025 I made a sixty gram batch with dried orange peel at what I intended to be five percent. I had a new bag of peel, cut much finer than the previous one, and I weighed three grams of it as usual. The resulting blend was undrinkable by the middle of the second week: sharp, perfumed, and with the peel completely covering a black leaf I had chosen specifically because I liked it on its own.
The weight was right and the blend was wrong, which took me a while to accept. Three grams of coarse peel in roughly centimetre pieces has a fraction of the surface area of three grams cut to two or three millimetres, and surface area is what does the extracting in a two minute steep. So weight alone is not sufficient for anything aromatic where the cut can vary. What I do now is note the cut in the entry, roughly, in millimetres, and treat a change of cut as a reason to remake at a lower percentage and taste before committing the rest.
I salvaged about two thirds of it by blending the ruined batch back into a further ninety grams of the plain black leaf, which brought the peel down to a shade under two percent and produced something perfectly pleasant that I drank through February. That was luck rather than a technique. The forty grams I could not rescue went into the bin and into the notebook, which is where it has been more use.