Some plants have extremely sensitive biological thermometers.
Conclusion (So)
If your thermometer shows 1°, and you're seeing uncurled rhododendron leaves as well as crocuses with mature but unopened blossoms, then you know that thermometer is accurate to within +/- 1 °.
Evidence
Once it's 2° or hotter, mature crocus blossoms are open. (if they're closed it's less than 2° C)
Once it's below 0°, leaves of rhododendrons curl up. (if they're uncurled, it's 0° or greater)
Evaluate
If we're seeing closed crocus we know it's less that 2°, and if we're seeing uncurled rhododendron leaves, we know it's 0° or greater.
So we know the only eligible temperature range is 0° ≤ x < 2°
If the thermometer measures 1° and has an accuracy reading within plus or minus 1°, then it could be as cool as 0° and as warm as 2°.
So ... this conclusion seems very valid.
Goal
We'd probably be pretty stumped. It seems like the math works and that saying .
But there must be some way to resist that conclusion. So let's look for some answer that would allow us to say that even though we're seeing uncurled leaves and unopened blossoms and a thermometer reading of 1°, it's still possible that the thermometer is more than 1° away from the real temperature.