|MadSci Network: Physics|
There is a wonderful book entitled “The Flying Circus of Physics”, by Jearl Walker, published by John Wiley and Sons. I checked there and this is the answer Walker has to your question. “The constriction is sufficiently narrow for the mercury to pass it only under pressure, either from thermal expansion or from the “centrifugal” force when is is swung in an arc. When cooling, the mercury thread breaks at the constriction because the intermolecular forces in the mercury are not strong enough to pull the upper column of mercury through the constriction.” Walker includes several references. I would add to this that a liquid is strong under compression but weak under tension. In other words, while it is true that if you push on a liquid it will push back and resist being compressed, if you pull on a liquid it is possible to rip the liquid apart and form pockets of what is nearly a perfect vacuum. This process is known as cavitation. I suspect that your question about surface tension is relevant, because at some point, before the thread of mercury is completely broken, a small bubble most form somewhere (In the middle of the thread or at the wall?). I am certain that the energetics involved in the formation of this original, small bubble must depend on surface tension. I would guess that liquids with higher surface tension would require larger negative pressures in order to cavitate. A brief word about surface tension. One way to think about this is to suppose you could form a thin liquid film between two identical circular metal wire rings that are close to one another and concentirc (centers lined up). To make this experiment simple to think about, imagine performing it in the space shuttle, while in orbit. Then there are no gravitational forces to consider. You would find that the two rings are being pulled together with a force proportional to the circumference of the circular rings, and that different liquids would pull with differing forces. This force is surface tension. It has the units of force/length. Here is an experiment you might try concerning the fever thermometer. Is it possible to cool the thermomether down very, very slowly so that the mercury thread does not break? You may have to be very clever to think of a way to do this but I’m betting it can be done, thought it might take a very long time (tens of minutes, a few hours?). I hope this has been of some help.
Try the links in the MadSci Library for more information on Physics.