To be called only for Monte Carlo datasets produced before December 1993; for more recent datasets one has to use QDEDX (see the use of the CHTSIM function just below).
This routine serves the same purpose as QDEDX, but treats Monte Carlo differently. QDEDX takes the dE/dx from the detailed simulation program TPCSIM. QDEDXM, however, only takes the number of samples and the track length from TPCSIM, from which a prediction for the resolution is obtained. The measured momentum and the Monte Carlo true mass then are used to predict the mean dE/dx, which is smeared by a gaussian random number to give the simulated dE/dx.
The advantage of this approach is that it is easy to adjust the parameterization to give agreement with data, whereas to do so with TPCSIM is nontrivial and would require regeneration of the Monte Carlo data set. The disadvantage is that the non-gaussian tails (which are small and arise primarily on the high side, due to unresolved track overlap) are not simulated. An option does exist to try to get the best of both worlds: by calling QMTAIL one can set a parameter to tell QDEDXM to retain the tail simulated by TPCSIM beyond a specified number of standard deviations. The distribution below that number of standard deviations then is obtained from the gaussian random number generator. Clearly this solution is not perfect, since the distributions generally will not match at the chosen cut value.
The arguments to this routine are identical to those of QDEDX. When QDEDXM is used on Monte Carlo events, error code 6 means that no Monte Carlo truth information is available. Note that if QDEDXM is used with real data, it is identical to QDEDX.
QMTAIL is an entry point in QDEDXM which can be used to set the cut value, in standard deviations, beyond which the dE/dx non-gaussian tail produced by TPCSIM is retained. By default, CUT is set to 999.