Usually you leave a reaction uncovered to induce crystallization, because this allows the solvent to evaporate. As solvent evaporates, the solubility of the product decreases, causing it to exit solution and begin to nucleate crystals. Once nucleation has occurred, crystals precipitate more readily.
Whether partial or totally uncovered isn't always important depending on how volatile the solvent is or how saturated the solution is. Partially covered will slow the rate, especially if the solvent isn't volatile, but it can block loose particles from falling in. On the other hand, a saturated solution doesn't need much evaporation to go below the solubility limit of your product and instigate nucleation. Whatever your case may be, a watchglass is not a sturdy seal against evaporation so it seems like your product was still able to crystallize.
Also, depending on the heat of the reaction and melting point of the product, sometimes evaporation isn't necessary at all, but it's still convenient to allow as much solvent as possible to evaporate in order to isolate dry crystals. If your solvent dries over night then you don't have to remove it the next day, so as a rule of thumb a reaction product is left uncovered unless there's something air/moisture sensitive present.