Last updated 18-Oct-1999
A couple of major errors have been discovered in my thesis.
One is that Tables D.16 and
D.17 are incorrect -- in fact, they were incorrect three times!
Another is an error in the column heading and in the values of
one column in both Tables D.4 and D.5.
There are also a
few typos and other miscellaneous errors. In the interests of accuracy
and ego, here are most of the known mistakes in the thesis:
- The poem, Cosmic
Gall by John Updike, is slightly mis-quoted. The third line should
read And do not interact at all. Janet Conrad changed "do not"
to "hardly" on one of her transparencies, and I copied that text into
my thesis. Also, there should be a comma after the word "And" on the
12th line: And, scorning barriers of class...
The lesson: always go back to the source material!
- Page 7
- Equation 1.5: It should be (k - k'), not (k + k').
- Page 9
- Equation 1.13: The anti-neutrino subscript should be a neutrino subscript in both
this equation and the following paragraph.
- Page 11
- Equation 1.22: The sign in the parenthetical term is wrong. It should be
1 + nu2/Q2.
- Page 11
- Equation 1.23: Remove Emu2, put a "nu" before W2,
and put an "M" before W1.
- Page 12
- Equation 1.28: The sign in the parenthetical term is wrong. It should be
1 + Q2/(2Mx)2.
- Pages 11 and 12
- Equations 1.23, 1.27, and 1.29 have the same mistake: The y-coefficient
of the last term should be y(1-y/2). It presently reads y(1-y2/2).
- Page 14
- Beginning paragraph, third sentence: "...they travel collinearly..."
- Page 28
- It was pointed out to me that it is useful to quote
the range of the Q2 variation used to determine the error
due to the normalization and factorization scales. In Ref. [Vir92], k
was varied between 0.1 and 4.0 to determine the shift in alpha_s,
where the new Q2 scale was taken to be k*Q2. k was
varied separately in the normalization and factorization scale analyses.
- Page 46
- Second bulleted item: The SUPERLOs are attenuated by a factor of six, not a
factor of ten.
- Page 92
- Six lines down from top of page: the semi-colon should be a period.
- Pages 89, 98, and 116
- Equations 4.1, 4.13, and 4.80 all have the same error: We do not correct to a
massless propagator (W mass = 0), we correct to an infinite-mass propagator. (Note:
in the FORTRAN code, this is accomplished by not executing the propagator-effect
code when we want to compute the correction.)
- Page 113
- Equation 4.55: The minus sign should be a plus sign:
ksi = x*(1 + mc2/Q2)
- Page 116
- Equation 4.79: The second "squared" in the denominator should be outside the
parenthesis. That is, the denominator should be
- Page 159
- Near the top of the page, xF3 is the average of xF3(nu) and
The "+" sign in the middle line of Equation 6.9 should be a "-" sign.
- Page 169
- It is Cynthia McNulty who's working on the alternate method of
extracting structure functions.
- Page 211
- Equation 7.21: The coefficient of the first term should be
- Page 215
- Figure 7.3: The presentation of the xF3 higher-twist (HT) results from Sidorov is incorrect.
The D&W HT values (which are denoted by D2) are determined by
xF3(HT) = xF3(QCD) * (1 + D2/Q2)
whereas the values published by Sidorov come from:
xF3(HT) = xF3(QCD) + D2/Q2
However, the conclusion of the comparison between D&W and the Sidorov measurement
would not change even if the Figure were corrected: the D&W HT calculation is
about twice as big as the measured HT correction.
- Page 221
7.6 and 7.7 show a variation of Lambda_MS as a function of x and
Q2 cuts. At the time that the thesis was written, I had no
explanation for the variation.
Kataev, Parente, and Sidorov performed a NNLO QCD analysis of the CCFR data to
higher orders of alpha_s; their results are in
Among other results, they show that the variation shown in Tables 7.6 and 7.7 goes away if
the NNLO and NNNLO QCD terms are included in the fit.
This is not really an error, but a recognition of new physics results.
- Page 223
- A theoretical error is missing from the error discussion: The value of
Lambda_MSbar is almost unchanged by whether or not nuclear corrections
are applied to the structure functions (see Section 22.214.171.124 for a discussion
of the nuclear corrections). A nuclear correction affects the shape of the
parton distribution functions, but not the slopes of the structure functions.
- Page 244
- Table 7.18: The values of Lambda for the Basic Fit line are
incorrect for both Combined Error columns. The value of Lambda for the
Combined Error is the same as the value in the
corresponding Stat and Syst column.
Note for second paragraph on the page:
See Section 126.96.36.199 (page 27) for a discussion of the QCD renormalization and
- Page 251
- In the caption to Figure 8.5, the reference to the SLAC E87 data is incorrect.
It should have been: Bodek et al., PRL 50:1431 (1983).
- Pages 339 through 346
- Tables D.4 and D.5 have two major errors.
The first error is that the columns of the table are incorrectly labeled.
The correct labels should be (reading across the top of the table, and
allowing for the typesetting limitations of HTML):
delta F2 (stat)
B/A nu 744
B/A nu 770
B/A nubar 744
B/A nubar 770
for D.4, and the corresponding xF3 labels for Table D.5.
The second error is that values of the column labeled "B/A nubar 744" (which
should have been labeled "B/A nu 770") are incorrect (thanks to Michiel Botje
for spotting that the values were the same as those in the previous column).
The correct flux systematic error tables with correct column headings and the correct values
are in the
SF final results area for both
- Pages 387 through 394
- Tables D.16 and D.17 are incorrect. There were actually three mistakes on my part; the first was
spotted by Arie Bodek, the second by Csaba Boros, and the third by Un-Ki Yang.
If you're going to use these tables, make
sure you get the ones created on or after 27-Sep-1999.
The correct values for removing the
physics model corrections used in the analysis are in the SF final results area
and are available both for F2
- xF3 sea corrections table
- At the request of Stefan Reiss at H1, a new corrections
has been added to the FTP site (and this was also corrected on 27-Sep-1999). It gives the ratio:
xF3 (no s-c correction) / xF3 (with s-c correction)
There is a separate correction listed for neutrinos and anti-neutrinos. In theory, if you
multiply the published xF3 by the columns in this table, you'll get xF3(neutrino) and
xF3(anti-neutrino) separately. In practice, what you'll get is the strange sea as
measured in the leading-order CCFR dimuon analysis. (Yes, the negative numbers in the anti-neutrino
ratio are correct: at low x and low Q2 xF3(anti-neutrino) is negative.)
- Page 440
- In Figure G.4, the curve for the CCFR NLO dimuon analysis is
incorrect. This was due to an oversight in Andrew Bazarko's thesis: in
Figure 7.27 of that thesis, the plot labeled Q2=10 is actually
for Q2=4. The plot
in the ftp site for my thesis has been corrected and should be used
instead of my Figure G.4.
- Pages 438-440
- It has been noted that the CCFR parton distribution functions shown in these
plots are for an iron target, while the global fits are calculated for a proton
target. No nuclear correction was applied to the CCFR PDFs for these plots.
Send e-mail to Bill Seligman