Egon, I agree that the effects you mentioin exist but I don't agree that they are the reason for the results reported.
I'm still waiting for the delta T (in seconds) between the freezing times of two samples starting at the same temp. This sort of calibrates and proves the experimental setup capable of measuring the experiment of interest. If you get too much variation betwen equal temp samples then the experimental setup is flawed or there are things going on not accounted for or both.
Now to reduce the problem a bit... The data reported said that the originally hot sample caught up with the cold sample and passed it in loosing heat. This is supposed to support the contention that a warmer sample freezes first.
I claim that at the instant the originally warmer sample passed the originally cooler sample (in a negative direction thermometer wise) that it then became the cooler sample and the originally cooler sample is now the warmer sample. So the NOW warmer sample does not freeze first, does it!!!
There is something beyond simple calorimetry going on here and I am not claiming that it has to be experimental error. It could be all or in part but we just don't know yet.
What if you fill a large insulated coffee carafe with tap water and then fill one side of the tray from the carafe. Put the tray in the chamber for a few minutes and then remove it and fill the other side from the carafe and return the tray to the chamber. One sample is warmer than the other but both have similar disolved gas content.
If the experiment were repeated succsessively with a smaller and smaller temp differential between samples you would expect the freeze times to converge. Why should there be an inflection point above which the hot will cool down, catch the cool and pass it but below which they act the same.
I say again, this is interesting. The calorimetry is really just a yardstick against which the real mystery is able to be observed. Everyone who has a modicum of physical intuition knows that given a "fair" start the cold should freeze first. WHY it doesn't, not THAT it doesn't, is the interesting part that should be the focus of the investigation.
If masses of the samples are equal (not volumes) and the disolved air is equal from sample to sample then the results would be even more interesting. It might be interesting to start with triple distilled water that is heated to boiling and placed into two full containers and kept from contact with air. cool one container and reheat the other in a microwave to get a large delta temp and then put the respective samples in the tray and into the chamber and see what happens.
Triple distilled water is available by the quart or gallon real cheaply at most grocery stores.
CAUTION: Be careful using a microwave to heat a full container. The lid must allow pressure to escape. Any contaminate introduced into water heated beyond the boiling point but without boiling can cause a rapid boiling effect, steam explosion.
Pat
