The obvious setting is not the right one

Every hydraulic press has a number you can push it to. Ours will go past 2,000 psi comfortably, and at that setting it extracts noticeably more juice from the same weight of fruit. For the first few months, that is exactly what we did.

The problem showed up in the citrus. Sanbokan has a thick albedo — the white layer under the peel — and under high pressure that layer gives up compounds you do not want. The juice came out fractionally more bitter and, more annoyingly, the bitterness grew over the following hours.

What we tested

We ran the same lot at three settings across a fortnight: 1,400, 1,600, and 2,100 psi. Same fruit, same day, same press, same bottles. Then we tasted each at pressing, at four hours, and at twenty-four.

Yield behaved as expected. The high setting produced roughly a tenth more juice by volume than the low one. On yield alone the decision would have been trivial.

Taste did not follow yield. At pressing, all three were acceptable. At four hours the 2,100 sample had a distinct back-of-the-throat bitterness that the others did not. By twenty-four hours it was the only one we would not have served.

Where 1,600 came from

1,400 was clean but wasteful — we were leaving too much in the cake, and with six growers and small lots, waste is not something we can be relaxed about. 1,600 held the flavour and recovered most of the difference.

It is not a magic number and I would not expect it to transfer to a different press or different fruit. It is the number that suits this machine, this citrus, and the fact that we sell within a day.

Pears behave differently

Maizuru pears do not have the albedo problem, and they tolerate higher pressure without turning bitter. In principle we could run them hotter and switch settings between lots.

We do not, mostly for a boring reason: changing the setting between lots means recalibrating, and recalibrating mid-service is how mistakes happen. A single setting that is right for the harder case and merely fine for the easier one is the arrangement a small bar can actually execute.

What this means for the bottle

It means our yield per kilo is lower than it could be, and the price reflects that. It also means the bottle you open six hours after we pressed it tastes close to what came out of the press.

If you have ever bought cold-pressed juice that was pleasant at the counter and harsh by the afternoon, pressure is one of the likelier explanations. Not the only one — oxidation and temperature matter too — but the one nobody puts on the label.

The honest caveat

This was a two-week test in one bar with one press, judged by three people who all knew which sample was which. It is not a controlled study and I would not defend it as one.

It was enough to change what we do, which is the only standard it needed to meet. If someone runs it properly and finds the opposite, I would genuinely like to read it.