This is a lovely demonstration of the way water's density differs with temperature and convection currents (you don't actually see the currents, but you see the end result).
Allow a pitcher of blue-colored water to cool in the refrigerator overnight.
At demonstration time, prepare a pitcher of hot tap water. Color this water yellow.
Fill a jar (or cup) all the way with blue water. Fill an identical jar all the way with yellow water.
Place an index card on top of the blue jar. Carefully turn the jar over and set it on top of the yellow jar - make sure the rims line up.
Ask for hypotheses as to what will happen when you slide the card out. Slide the card out -- the blue water sinks, mixing with the yellow, creating green water.
Now try again....
Prepare the jars in the same way. But, this time, place the index card on the yellow jar, and place the yellow jar on top of the blue jar.
Remove the card and watch.....
You'll get a little green water right at the interface, but the yellow and blue water will mostly remain separate.
Why did you get two different results?
Cold water is denser than hot water - it sinks. When the cold water was on top of the warm, it sank to the bottom of the vessel, mixing with the warm, as evidenced by the mixing of colors.
When the cold water was on the bottom, it was content to stay right there. Just a little mixing occurs right where the two temperatures meet. What do you think would happen if you allowed it to sit for awhile? Would the colors remain separate, or would they eventually mix?