You can, because unless this is something you made, on the side of the graduated cylinder, it should list the error range, right near where it says at what temperature it was calibrated at. You would then be assuming that through wear and tear and age that it has not changed much.
A lot of the times, this is worn off so you can not read it, though.
I have never used a cylinder that large that does no have subdivision at the mL level. With the ones at the mL level, an average, based on my experience is ±1.3 mL. Without the mL divisions, it would be a safe bet at minimum to double that.
Ohh, and one more thing. The answer to your question is no, absolutely not. With a 500 mL graduated cylinder, there is no way you can say 537 ± x, as it does not hold that much
.