Changes made in the paper draft from: version 1 jpsi_fmu_prd.ps (4-dec-2001) to version 2 jpsi_fmu_prd.ps (6-mar-2002) based on comments received from two institutions in the collaboration: Purdue and Illinois. Page 1, Abstract Comment: There is a mistake in the data summation in Table 2. response: Total cross section number increased from 174 pb to 197 pb, and statistical and systematic errors increased. The cross section change followed a different treatment of the data in Table 2, p19, discussed below. Comment: Too many references to D0. response: The second specific reference to the D0 experiment has been removed. Page 1, Introduction Comment: c cbar states other than chi could contribute. response: The letter 'chi' has been removed from 'c cbar chi states'. Page 2, Detector No comment here. Change made by author. 'Spherical coordinates (r,theta,phi)' has been changed to 'Polar and azimuthal angles (theta,phi)' to correspond to the inset in Fig 1. Page 8, Forward Muon Trigger Comment: 1-1-1 trigger road jargon, and not defined. response: First paragraph 'Two basic cell patterns were designed ...' has been changed to define more clearly the trigger roads through the toroid drift chambers. Page 9, Luminosity Comment: curvy 'L' preferred for luminosity. Response: The symbol 'L' has been changed to script. Page 10, Selection Criteria Comment: Figs 7 and 8 should be combined. Response: done. 'The like sign background ..., but was only 8%' has been changed to 5%, and the numbers on the plots Figs 7 and 8 (which are now combined to Fig 7) have been corrected. The error was due to a feature of PAW, which prints the total number of events you give it, and not the total number plotted on the histogram. The cuts now match. This error did not affect the numbers in Table 2. Page 11 Comment: ideal chiprob should be defined better. Response: cut 5 'ideal chiprob' changed to 'ideal chisquared probability' Page 12 Data and Monte Carlo Comment: How much is signal and how much is background? response: a sentence has been added after '...fitted to a linear background plus a Gaussian signal.' to clarify that the full mass window was needed to fit the background. Figure 10 has become Fig 9 because Figs 7 and 8 have been combined. The wording of the sentence 'There were 1207 events...' has been changed to emphasize that Fig 9 has 1207 events in the peak between 2.0 and 4.4 GeV mass. Comment: Why not do the signal fraction calculation in each pt bin? response: OK The text at the bottom of p 12, which appears on page 16 of the new version, has been re-written to reflect a calculation of the signal fraction in each pt bin separately, rather than using an average value. Table 2 has been changed to incorporate this new method. Page 16 Table 1 errors have been added to columns 4 and 5. The numbering of the figures has changed because old figures 7 and 8 are now combined to fig 7. Comment: Why is a B->J/psi monte carlo sufficient, when this is not the only, or even dominant source of J/psi's? middle of page 16 'The detector acceptance for J/psi->mu+mu-...' has been expanded to make it clear that there are three independent variables in the problem. Two of them are adequately modeled, and the uncertainty in the third, the angular distribution in the J/psi rest frame, is a systematic error. Page 17 Results This section has been completely re-written. Table 2 now has the pt dependent signal fraction s calculated for each bin, with errors. Because s has been added as col 3 in Table 2, an extra column has been added to Table 3. The numbers in col 6 of the old Table 2 and col 2 of the new Table 3 are not identical, because of the pt dependent factor s. The cross section numbers in the final column of Table 3 are very close. The highest pt bin, which contains no events, has been dropped. Comment: The arithmetic in the cross section calculation is in error. Response: OK The total cross section is obtained by summing col 2 of Table 3, rather than summing the raw data as was done before.