Lawrence Berkeley National Laboratory

All too often we rely on Monte Carlo simulations without worrying too much about basic physics. It is possible to start with a very simple calorimeter (a big cylinder) and learn the functional form of pi e by aninduction argument. Monte Carlo simulations provide sanity checks and constants. A power-law functional form describes test beam results surprisingly well. The prediction that calorimeters respond differently to protons and pions of the same energy was unexpected. The effect was later demonstrated by the CMS forward calorimeter group, using the most noncompensating calorimeter ever built. Calorimeter resolution is dominated by fluctuations in piz production and the energy deposit by neutrons. The DREAM collaboration has recently used a dual readout calorimeter to eliminate the first of these. Ultimate resolution depends on measuring neutrons on an event-by-event basis as well.