Fishless cycling in the cold months
I set up a second filter to seed, in a 20-liter tub beside the main tank, with the idea of having a spare sponge ready if the main one ever failed. December in the kitchen is about 19 to 20 °C, and nobody warns you how much that slows things down. The whole process took 31 days. In July I expect it would have taken half that.
Ammonia source was plain household ammonia with no surfactant (it does not foam when shaken). I dosed to 2 mg/l on day one, checked by test kit, which took about 1.2 ml in 20 liters. Then I waited. Days one to seven: ammonia stayed at 2, nitrite zero. Day eight: nitrite finally registered at 0.25. Day ten: ammonia down to 1, nitrite 1. Day twelve: ammonia zero, nitrite off the top of the chart, and here I made the mistake. I redosed ammonia to 2 mg/l on top of a nitrite reading over 5, and the nitrite stayed pinned at the top of the chart for eleven days. Nitrite that high stalls the bacteria that should be eating it.
A 50 percent water change on day 23 brought nitrite to about 2, and from there it moved: day 26 nitrite 0.5, day 28 nitrite 0.25, day 31 ammonia and nitrite both zero within 24 hours of a 2 mg/l dose, which is the finish line. Nitrate at the end read about 40 mg/l, as it should.
Things that helped: a small heater borrowed from the main tank for the last ten days, which took the tub to 26 °C and visibly sped things up; and a squeeze of the established filter sponge into the tub on day one, which is probably why nitrite appeared on day eight rather than day fourteen.
Things that did not matter: pH. The tub sat at 7.4 the whole time and I never adjusted it. Light did not matter either; the tub lived under a towel. The spare sponge now sits in the main filter behind the active one, quietly staying alive, and the tub is back to holding potatoes.