The Silver Iodide Conundrum

While searching for preciptation reaction crystalisation patterns for my Scientific Rainbow Glitter (using possibly hazardous chemicals), I tested the reaction AgNO3(aq) + KI(aq) ==> AgI(s) + KNO3(aq). This reaction had some interesting results, to say the least. After first mixing it, everything semmed normal. It created a creamy white substance, and since KNO3 clear (proved by Golden Rain), AgI must have given it this colour. Therefore, putting it through filter paper should remove the AgI, and give you clear KNO3 in a test tube, and a white powder on the filter paper. This is where the anomalies began. Rather than this, the substance that drained through the filter paper was the same white-ish colour as before, leaving only silvery-grey traces behind on the filter paper. I was confused by this, so make sure it wasn't a fluke, I did the reaction again. This time, I noticed that rather than settling in the base of the test tube as a powder like other precipitates, it seemed to have dissolved into the solution. After filtering it once more, I discovered the same result, meaning it was not just a fluke. Something was amiss. Just in case, I conducted a third test, and once again, had the odd results. I recorded five possible chemicals that may have caused this to happening:

1. Potassium Nitrate

2. Silver Iodide

3. Silver

4. Iodide

5. The combination of Potassium Nitrate and Silver Iodide

I then met up with Mr. Bruce about it, and he added a sixth possible reason:

6. The sliver iodide particles are too small for the filter to catch, and they slip through into the solution

From the beginning, two of these could be crossed off the list: Potassium Nitrate and Iodide. Both of these were part of the "Golden Rain" experiment, and didn't cause any funny buisness. For the other threechemicals, I devised a way to test each:

1. Silver Iodide - Mix it with water

2. Silver - NaOH + AgCl ==> NaCl + AgOH 

3. The combination of Potassium Nitrate and Silver Iodide - If it is not anything else, it is probably this. 

I didn't have a way to test the filter paper theory, so I had to test the others first.