Lieber Werner,

Vielen Dank fuer die schnelle Antwort.

> In case of the total charm cross section, I had completely misunderstood what the paper was trying to 
> say. I found this out when discussing the issue with Bernhard. I wonder if it wouldn't be even clearer
> of one would write (p 7 col 1):
> "From the fractions ... is found to account for f_D = 89.1...%. Assuming the same fraction, the present study
> derives the charm cross section per nucleon to
> sigma(ccbar)=(sigma(D0)+sigma(D+)+sigma(Ds))/(2 f_D),
> where the factor 2 accounts for the charge-conjugated states which are included in the D cross sections." or similar
>
Excellent, I have built this in.

> On the pt distribution, I still have trouble figuring out what the description is tyring to 
> say. The problem is the "The distribution of background events N(pt^2)..." sentence. 
> (Unfortunately, Bernhard is away otherwise I could ask him). I would think, that one can either 
> measure the Xsect in some pt bins by sideband subtraction and then fit the pt-dependent Xsect 
> with some shape. Then, one doesn't need a parametrization for the pt-dependence of the background. 
> Or one does some kind of global fit, assuming a shape of the pt-dependence of the BG and of the 
> signal. In the latter case, one doesn't get data points versus pt (as shown in the figure), but 
> only some fit parameters. ... Did I make my problem clear - somehow I seem to misunderstand something??
>

Indeed, this part of the paper was not clear enough. We have added a paragraph after the 
first paragraph in the section 5.2, which I hope makes this clear.

'The parameters of the measured differential cross sections were determined in the following 
way. Because of the low statistics in individual  bins of $\pTsq$ and $\xF$, a
direct fit to the data points of Fig.~\ref{dsigma_dpt_dxf.eps}
is not adequate. Instead, a simultaneous binned likelihood fit of the $\pTsq$ and $\xF$
 distributions for events from the mass signal window and from the sideband windows was 
performed, by which the poissonian errors could be treated correctly.'




> Concerning Fig. 5: I suggest to mention explicitly that the normalization is fitted for each 
> species independently. Also, it is not clear (to me) if the HERA-B point is included in the fit, 
> or if the fit is seen as a summary of previous data. Maybe mention that.
>
Normalisation: I will add this. HERA-B points: are included, and we also say this in the paragraph 
where Fig 5 is discussed.

Gruss
Peter