Are natural laws ever violated?

Apparent 'violations' of natural laws are easy to contrive if you ignore the restrictions on the data that the law is based on.

For example, Charles' law states that volume is proportional to temperature for a fixed number of moles of ideal gas at fixed pressure. Charles' law is usually applied as an equation:

Vinitial/Tinitial = Vfinal/Tfinal

Notice that the restriction isn't included in the equation. Unfortunately, that causes some people forget all about restrictions when applying a law in equation form. Could you safely use Charles' Law to predict the temperature of the CO2 trapped in the neck of a soda bottle, after you pop the top, given its initial temperature and its initial and final volumes? Could you use Charles' Law to predict the final volume of 1L of water heated from 0°C to 100°C?

When experimental fact is in conflict with a law, the law must be either corrected or discarded. This is a rare occurrence, because laws must summarize a large body of experimental data before they're elevated to law status. But it does happen. For example:

Until 1956 all evidence indicated that processes involving elementary particles would be essentially unchanged when the directions of the particles were replaced by their mirror-image directions. This "law of conservation of parity" was disproved when an elegant experiment performed by physicists at the National Bureau of Standards (now the National Institute for Standards Technology, NIST) showed that the mirror-images of a process were in fact distinguishable. See The Fall of Parity at NIST's web site for an account of the experiment. Physicists had no experimental evidence that reversing time would change the physics of elementary particles. For example, small particles called 'kaons' can transform into 'antikaons', and vice versa. If conditions are carefully controlled, one expects the rate of kaon-to-antikaon transformation to be identical to the antikaon-to-kaon rate, because one process is just the reverse of another. In a landmark experiment in 1998, physicists at the European Laboratory for Particle Physics (CERN) and the Fermi National Accelerator Laboratory (Fermilab) have observed differences in the forward and reverse processes, which shows that there is a distinction between moving forward and backward in time.