1: \documentclass[superscriptaddress,showkeys,showpacs]{revtex4}
2: \usepackage{graphicx}
3: \usepackage[tbtags]{amsmath}
4: \begin{document}
5: \title{Sea quark polarization and semi-inclusive DIS data.}
6:
7: \author{X. Jiang}
8: \email{jiang@jlab.org}
9: \affiliation{Department of Physics and Astronomy, Rutgers, the State
10: University of New Jersey\\ 136 Frelinghuysen Road, Piscataway, NJ 08854 USA.
11: }
12: \author{G. A. Navarro}
13: \email{gabin@df.uba.ar}
14: \affiliation{Departamento de F\'{\i}sica,
15: Universidad de Buenos Aires\\ Ciudad Universitaria, Pab.1 (1428)
16: Buenos Aires, Argentina}
17: \author{R. Sassot}
18: \email{sassot@df.uba.ar}
19: \affiliation{Departamento de F\'{\i}sica,
20: Universidad de Buenos Aires\\ Ciudad Universitaria, Pab.1 (1428)
21: Buenos Aires, Argentina}
22: \begin{abstract}
23: We investigate the potential impact of forthcoming Jefferson Lab semi-inclusive
24: polarized deep inelastic scattering proton measurements in the determination
25: of the sea quark polarization in the nucleon by means of a next to leading
26: order global QCD analysis. Specifically, we estimate the resulting improvement
27: in the constraints on polarized parton densities for the different flavors,
28: which is found to be significant for up and strange quarks, and the
29: correlation between remaining uncertainty ranges for each of the parton
30: species.
31:
32:
33: \end{abstract}
34:
35: \pacs{12.38.Bx, 13.85.Ni}
36: \keywords{Semi-Inclusive DIS; perturbative QCD.}
37:
38: \maketitle
39:
40: \section{Introduction}
41:
42: The way in which sea quarks are polarized inside nucleons has been
43: a persisting question ever since the spin structure of the proton
44: began to be unveiled by polarized DIS experiments and even today, in spite of
45: several successful experimental programs, remains to a large extent unanswered.
46: Contrary to the common belief before the paradigmatic EMC experiment at the
47: end of 1980s \cite{Ashman:1987hv,Ashman:1989ig}, the data obtained
48: by the collaboration suggested that sea quarks and gluons in the nucleon
49: carried non negligible polarization. However this conclusion was, and has
50: been for many years, conditional upon rather strong assumptions on isospin
51: symmetry extended to polarized phenomena. In the subsequent years, isospin
52: symmetry itself become seriously questioned \cite{Kumano:1997cy} and
53: consequently the sea quark polarization turned into an even more elusive
54: question.
55:
56: Global QCD analyses including semi-inclusive measurements of polarized
57: lepton-nucleon deep inelastic processes began to change this situation
58: more recently \cite{deFlorian:1997ie,deFlorian:2000bm}, and today these
59: data allow to constrain the extraction of polarized parton densities in
60: QCD global fits \cite{deFlorian:2005mw}. The effectiveness of these
61: constraints of course relies on the precision of the data and this is why
62: the forthcoming generation of semi-inclusive experiments is crucial.
63:
64: In a QCD global fit the uncertainty range of the resulting parton densities
65: can be estimated by analyzing the profile of the $\chi^2$-function of the fit
66: to data against variations in the different features of the densities. This
67: technique has been widely used in extractions of unpolarized parton densities
68: \cite{martin:2003} and more recently has been implemented in the polarized case \cite{deFlorian:2005mw} providing reliable constraints on their different
69: features, such as the net polarization carried by each parton flavor.
70:
71: Of course, the values each flavor polarization, or some other parameter, may
72: take within these constraints are not independent, but become correlated.
73: Even in the case where there are enough independent observables to extract
74: in principle all the parton densities, the uncertainties in the measurements
75: of those observables, together with the theoretical uncertainties inherent
76: in the fitting procedure, conspire against the independence of parton densities
77: and results in correlations. A strong correlation between two parton densities,
78: consequently means that neither of them is actually well determined. The
79: inclusion of new and more precise data should not only reduce the uncertainty
80: ranges for each flavor but also those correlations.
81:
82: In this article we investigate the impact of the inclusion of a
83: series of semi-inclusive deep inelastic scattering measurements
84: to be performed at Jefferson Lab \cite{xiaodong} in a next to leading order
85: (NLO) QCD global fit to all the available inclusive and semi-inclusive deep
86: inelastic scattering data. In order to do this we take into account
87: the kinematical coverage, statistics and level of uncertainty
88: expected for the measurements. We also analyze the correlation
89: between the uncertainty ranges of the different sea quark
90: polarizations. As result of this analysis we find that the forthcoming
91: Jefferson Lab experiment will effectively contribute to constrain the
92: sea quark
93: polarization in the proton. The most significant improvement is found
94: in up sea quark distributions, and also with a noticeable effect for
95: strange quarks. The improvement in both distributions will be
96: related to the inclusion in the fit of more precise data on charged meson
97: electroproduction. Another interesting effect of the inclusion of
98: these data is the softening of the correlation between the allowed range for
99: the polarization of up and down sea quarks.
100:
101: In the following section, we settle definitions and conventions
102: for the global fitting procedure and the way in which we study the
103: profile of the $\chi^2$ function, we discuss the characteristics
104: of the forthcoming semi-inclusive experiment, and explain how we evaluate
105: the impact of it in a global fit. Then, we compare the results
106: coming from the analysis of the set of data available at present, against
107: those that would come from the data set enlarged with the forthcoming
108: measurements, both for the individual uncertainty
109: ranges, for the net polarization of the different flavors, and for the
110: correlations between flavors. Finally, we present our conclusions.
111:
112:
113: \section{Global QCD fits and new data}
114:
115: In the present analysis we implement the NLO QCD global fit to existing
116: data along the lines of what was done in reference
117: \cite{deFlorian:2005mw} but restricting the input fragmentation functions to
118: those of reference \cite{kretzer}, which were shown to give the best fits
119: to combined polarized data. The NLO expressions for both inclusive and
120: semi-inclusive spin-dependent asymmetries and evolution equations for the
121: parton densities can be found in \cite{inclusiva,newgr} and \cite{NPB}
122: respectively.
123:
124: The data sets analyzed include only points with $Q^2>1$ GeV$^2$,
125: listed in Table \ref{tab:table1}, and totaling 137, 139, and 37 points, from
126: proton, deuteron, and helium targets respectively, from polarized inclusive
127: deep inelastic scattering plus 60, 87, and
128: 18, from proton, deuteron, and helium targets respectively from semi-inclusive
129: deep inelastic scattering.
130:
131:
132:
133:
134: The main conclusions reached in reference
135: \cite{deFlorian:2005mw} were that using the Lagrange multiplier approach
136: \cite{Stump:2001gu} as a mean to explore the
137: profile of the $\chi ^2$ function against different degrees of polarization
138: in each parton flavor, definite estimates for the uncertainty in the net
139: polarization
140: of each flavor, and in the parameters of the polarized parton distributions,
141: can be obtained.
142: The overall result is a well constrained scenario where semi-inclusive data
143: is not only consistent with inclusive measurements, but improves the
144: constraining power of the fit for all the distributions, being crucial for
145: the light sea quarks.
146:
147: The best fits suggest an overall picture for the quark
148: densities at $10$ GeV$^2$ where, within uncertainties, up quarks are almost
149: $100\%$ polarized parallel to the proton, down quarks anti-parallel in a
150: similar proportion, and sea quarks have a small and flavor symmetric negative
151: polarization. The first moment of the gluon distribution is found to be in
152: agreement with the most recent direct measurements \cite{Ellis:2005cy}
153: close to 0.6, constrained to be smaller than 0.8 and larger than -0.05
154: within a conservative increase in the $\chi ^2$ value within a two percent
155: range ($\Delta \chi^2=2\%$).
156:
157: \begin{table}
158: \caption{\label{tab:table1} Inclusive and semi-inclusive data used in the fit.}
159: \begin{ruledtabular}
160: \begin{tabular}{ccccc}
161: Collaboration & Target& Final state & \# points & Refs. \\ \hline
162: EMC & proton& inclusive & 10 & \cite{Ashman:1989ig} \\
163: SMC & proton, deuteron & inclusive & 12, 12 & \cite{SMCi} \\
164: E-143 & proton, deuteron & inclusive & 82, 82 & \cite{E143} \\
165: E-155 & proton, deuteron & inclusive & 24, 24 & \cite{E155} \\
166: Hermes & proton,deuteron,helium& inclusive & 9, 9, 9 & \cite{HERMES} \\
167: E-142 & helium& inclusive & 8 & \cite{E142} \\
168: E-154 & helium& inclusive & 17 & \cite{E143} \\
169: Hall A & helium & inclusive & 3 & \cite{HALLA} \\
170: COMPASS & deuteron & inclusive & 12 & \cite{COMPASS} \\
171: \hline
172: SMC & proton,deuteron& $h^+$, $h^-$ & 24, 24 & \cite{SMC} \\
173: Hermes & proton, deuteron, helium & $h^+$, $h^-$, $\pi^+$, $\pi^-$, $K^+$, $K^-$, $K^T$ & 36,63,18
174: & \cite{HERMES} \\ \hline
175: \multicolumn{3}{c}{Total} & 478 & \\
176: \end{tabular}
177: \end{ruledtabular}
178: \end{table}
179:
180:
181:
182: In order to evaluate the impact of the forthcoming semi-inclusive proton data of
183: Jefferson Lab, we included in the global analysis the expected values of the $\pi^+$, $\pi^-$, $K^+$ and $K^-$ semi-inclusive asymmetries on a polarized proton target,
184: computed with the
185: best set of parton densities obtained in \cite{deFlorian:2005mw} with expected
186: experimental uncertainties, as an additional set of points to be fitted.
187: The projected statistical accuracies of these asymmetries are based on a total of 225 hours of 6 GeV polarized electron beam on a polarized $NH_3$ target. The electron beam current is assumed to be 80 nA with a polarization of 80 \%. The standard Jefferson Lab Hall C polarized $NH_3$ target of 3 cm thickness and 80 \% polarization is assumed. The scattered electron will be detected at 30 degree with an array of lead-glass detectors in conjunction with a threshold gas Cherenkov counter, covering a solid angle of 210 msr. The produced hadron will be detected in coincidence using the standard Hall C High Momentum Spectrometer (HMS) at 10.8 degree and a central momentum of 2.7 GeV/c ($z_{\pi} \approx 0.5$). The HMS spectrometer has a solid angle of 6 msr and a momentum acceptance of $\pm 10 \%$.
188:
189: With this enlarged set of asymmetries to be fitted, we have redone the
190: analysis of \cite{deFlorian:2005mw} adding a detailed study of the
191: correlations, and compared the resulting constraints on
192: polarization with those of the original fit.
193:
194: \section{Results}
195:
196: \setlength{\unitlength}{1.mm}
197: \begin{figure}[b]
198: \includegraphics[width=12cm]{dist.eps}
199: \caption{Profile of the $\chi^2$ function against parton polarizations.}
200: \label{fig:par}
201: \end{figure}
202:
203:
204:
205:
206:
207: We start with the estimates for the uncertainties in the polarization of the
208: different quark flavors. In Figure \ref{fig:par} we show the outcome of varying
209: the $\chi^2$ of the NLO fits with the set of data available at present, in the
210: following referred to as ``standard fit'', and the ``improved fit'', which
211: includes the asymmetries expected to be measured by the E04-113, against the
212: first moment of the respective polarized parton densities
213: $\delta \overline{q}$ at $Q^2=10$ GeV$^2$, one at a time. This is, to minimize
214: \begin{equation}
215: \Phi(\lambda_q, a_j)=\chi^2(a_j)+\lambda_q\, \delta q(a_j) \,\,\,\,\,\,\,\,\,\,\,\, q=u,\overline{u},d,\overline{d},s,g.
216: \end{equation}
217: where $\lambda_q$ is the Lagrange multiplier associated to the
218: polarization of a given quark flavor $q$, $a_j$ are the parameters to be
219: fitted, and the $\chi^2$ definition is the most simple and commonly used in
220: fits to polarized data, namely,
221: \begin{equation}
222: \chi^2=\sum_{i=1}^N \frac{(T_i(a_j)-E_i)^2}{\sigma_i^2}\,.
223: \end{equation}
224: \setlength{\unitlength}{1.mm}
225: \begin{figure}[hbt]
226: \includegraphics[width=14cm]{bandas.eps}
227: \caption{Uncertainty bands for $A_p^{\pi +}$ and $A_p^{\pi -}$ asymmetries
228: coming the ``improved fit'' and E04-113 expected uncertainties.}
229: \label{fig:new}
230: \end{figure}
231: In Equation (2), $E_i$ is the measured value of a given observable,
232: $T_i$ is the
233: corresponding theoretical estimate computed with a given set of parameters
234: for the polarized parton densities, and $\sigma_i$ is the error associated
235: with the measurement, usually taken to be the addition of the reported
236: statistical and systematic errors in quadrature. Notice that the additional
237: set of asymmetries included does not contribute to $\chi^2$ when it is computed with
238: parton densities corresponding to the best fit of reference
239: \cite{deFlorian:2005mw}, the same densities used to generate the asymmetries,
240: situation that happens at the minima of the curves. As the distributions
241: change in
242: order to increase or reduce the polarization of a given flavor, the $\chi^2$
243: obtained with one or another set begin to differ. The solid lines in
244: Figure \ref{fig:par} correspond to the analysis of the standard set of data,
245: while the dashed lines includes the estimated impact of future measurements.
246: \setlength{\unitlength}{1.mm}
247: \begin{figure}[hbt]
248: \includegraphics[width=19cm]{elipses5.eps}
249: \caption{Correlations between the flavor polarizations within a 5\% increase
250: in $\chi^2$ in both the standard and the improved fits.}
251: \label{fig:eli}
252: \end{figure}
253:
254: As expected, the most noticeable effect is in $\overline{u}$ polarization
255: due to the maximal sensitivity of the semi-inclusive asymmetries on proton
256: targets to this distribution, as discussed in \cite{deFlorian:2000bm}.
257: For the $\overline{d}$ and $\overline{s}$ distributions, the impact of the
258: future measurements is comparatively suppressed by the weakness of the
259: electric charge factor associated with these flavors, however there is a
260: noticeable improvement for the $\overline{s}$ quark uncertainty near
261: the minimum. In the previous analysis this distribution was mainly constrained
262: by positivity resulting in flat $\chi^2$ distribution around the minimum but
263: here shows a nice parabolic profile. Notice that both in the
264: analysis of reference \cite{deFlorian:2000bm} and in the present one, we are
265: forced to assume $\Delta \overline{s}=\Delta s$ since there is not enough
266: data yet to discern alternatives. This assumption implies a strong constraint
267: on $\delta \overline{s}$ even though the measured asymmetries are less
268: sensitive to this distribution than to $\delta \overline{u}$ and
269: $\delta \overline{d}$. On the contrary, the relation between
270: $\delta \overline{u}$ and $\delta {u}$ and the same for $\delta \overline{d}$
271: and $\delta {d}$ comes from the fit. The impact on the gluon
272: distribution is not significant and mainly indirect, coming as found in
273: reference \cite{deFlorian:2000bm} through the constraints on the sea quark
274: distributions, which are now better defined. It is worth mentioning that the
275: impact of the kaon data is very mild, being mostly the pion asymmetries
276: responsible for the changes.
277:
278: As discussed in references \cite{deFlorian:2000bm,martin:2003}, in modern
279: extractions of parton densities it is customary to consider alternative sets
280: of parton densities within an increase between 2\% and 5\% in $\chi^2$, as a
281: conservative estimate for the range of uncertainty of the global fit. In order
282: to estimate the corresponding uncertainty range in the computation of a given
283: observable, it is customary to take it as the range of variation of the
284: observable within the alternative sets. This is precisely what we show in
285: Figure \ref{fig:new} where we plot the uncertainty
286: bands of $A_p^{\pi +}$ and $A_p^{\pi -}$ corresponding to
287: $\Delta \chi^2 = 5 \%$ in the improved fits as the area between the dashed
288: lines.
289: For comparison we include in the plots the values for these pion asymmetries
290: at the kinematics of the forthcoming Jefferson Lab experiment, computed with
291: the set of \cite{deFlorian:2005mw} and which where included as ``data'' in the
292: ``improved fit'', together with the expected error bars. In this way, we can
293: see not only the consistency of the results but also the appropriateness of
294: the choice of $\Delta \chi^2 = 5 \%$
295: Similar uncertainty bands obtained for $\Delta \chi^2 = 5 \%$ but in the
296: ``standard fit'' were found to be twice as large in \cite{deFlorian:2005mw}.
297:
298:
299: As it was pointed out in the introduction, an important feature to keep in
300: mind regarding the range of variation of the polarization of the different
301: flavors is that they are correlated. For example, they cannot be expected
302: to hold simultaneously; in consequence, for a given allowed range in
303: $\chi^2$, two or
304: more flavors may cannot take their respective maximum departures from the
305: best fit value together. For this reason it is worth while to study such
306: correlations between the different flavors, and how these correlations
307: change with the inclusion of additional data.
308: This can be done systematically generalizing Equation (1) for more
309: than one flavor polarization, with independent Lagrange multipliers,
310: scanning the profile of the $\chi^2$ function in the range of variation of
311: them.
312:
313:
314:
315: In Figure \ref{fig:eli} we show the allowed range of polarization within a 5\%
316: increase in $\chi^2$ in the $\delta \overline{u}-\delta \overline{d}$,
317: $\delta \overline{u}-\delta \overline{s}$, $\delta \overline{u}-\delta g$,
318: $\delta \overline{d}-\delta \overline{s}$, $\delta \overline{d}-\delta g$, and
319: $\delta \overline{s}-\delta g$ planes. In order to simplify the plots we have
320: approximated the actual contours by ellipses, the darker ones obtained with
321: the ``improved fit'', while lighter being those coming from the
322: ``standard fit''. Again the most prominent effect is the shrinkage of
323: the $\delta \overline{u}$ and $\delta \overline{s}$ uncertainty range.
324:
325: In these plots the correlations between the polarization of the different
326: flavors are represented by the angles between the axes of the ellipses and the
327: coordinated axes. A positive or negative $\pi/4$ difference would imply a
328: maximal positive or negative correlation respectively, and that both
329: polarizations are weakly constrained. This is the case, for example for
330: $\delta \overline{u}$ and $\delta \overline{d}$ in the ``standard fit'',
331: situation that is corrected in the improved version. In the remaining cases,
332: the axes of the ellipses are almost parallel to those of the coordinates
333: suggesting mild correlations between the different pairs of flavors.
334:
335: There is however, a remaining subtle correlation between
336: $\delta\overline{u}$ and $\delta\overline{d}$ and between
337: $\delta \overline{u}$ and $\delta\overline{s}$, that the enlarged set of
338: asymmetries included in the fit is still not able to remove.
339: In the first case the residual correlation is positive, while in the latter
340: it is negative.
341: The gluon polarization also seams to have negligible correlation with that
342: of the anti-quarks, with only a very slight positive tendency with
343: $\delta\overline{s}$ in the standard fit, which is removed in the improved
344: one.
345:
346:
347: \setlength{\unitlength}{1.mm}
348: \begin{figure}[hbt]
349: \includegraphics[width=9cm]{elipseseatot.eps}
350: \caption{Correlation between $\delta \Sigma_{\overline{q}}$ and $\delta g$.}
351: \label{fig:seatot}
352: \end{figure}
353:
354:
355:
356: Another interesting correlation to investigate is the one between the total
357: polarization carried by anti-quarks and that of the gluons. This in practice
358: amounts to associate a Lagrange multiplier in Equation (1) to the sum over
359: the anti-quark species $\delta \Sigma_{\overline{q}}=(\delta \overline{u}+
360: \delta \overline{d}+\delta \overline{s})$ and another to the gluon
361: polarization.
362: In Figure \ref{fig:seatot} we see that there is no significant correlation
363: between $\delta \Sigma_{\overline{q}}$ and $\delta g$. Notice also how
364: significantly the asymmetries expected to be measured by E04-113 will help
365: to constrain the anti-quark polarization.
366:
367:
368:
369:
370:
371: \section{Conclusions}
372: We have analyzed the potential impact of forthcoming semi-inclusive
373: polarized deep inelastic scattering proton measurements in the determination of
374: sea quark polarization in the nucleon by means of a next to leading order
375: global QCD analysis. We find that the inclusion of this data
376: will effectively contribute to constrain the sea quark
377: polarization in the proton. The most significant improvement is found
378: in the up anti-quark distribution, and with an also noticeable effect for
379: strange anti-quarks. For down anti-quarks, the new data will have a smaller
380: but nonnegligible effect.
381: Regarding the correlations, we found that the forthcoming data will reduce
382: the apparent correlation found between $\delta \overline{u}$ and $\delta
383: \overline{d}$ in the standard fit of reference \cite{deFlorian:2005mw}
384: leading to a picture where the sea quark densities and their uncertainties
385: can be determine independently.
386: \section{Acknowledgements}
387:
388: We warmly acknowledge Daniel de Florian for comments and suggestions.
389: R.S. is grateful to Jefferson LAB for the hospitality during his visit
390: where this analysis was completed. This work was partially supported by
391: CONICET, Fundaci\'on Antorchas, UBACYT and ANPCyT, Argentina and the
392: US National Science Foundation grant PHY 03-54871.
393:
394: \begin{thebibliography}{3}
395:
396: \bibitem{Ashman:1987hv}
397: J.~Ashman {\it et al.} [European Muon Collaboration],
398: Phys.\ Lett.\ B {\bf 206}, 364 (1988).
399: \bibitem{Ashman:1989ig}
400: J.~Ashman {\it et al.} [European Muon Collaboration],
401: Nucl.\ Phys.\ B {\bf 328}, 1 (1989).
402: \bibitem{Kumano:1997cy}
403: S.~Kumano,
404: Phys.\ Rept.\ {\bf 303}, 183 (1998).
405: \bibitem{deFlorian:1997ie}
406: D.~de Florian, O.~A.~Sampayo and R.~Sassot,
407: Phys.\ Rev.\ D {\bf 57}, 5803 (1998).
408: \bibitem{deFlorian:2000bm}
409: D.~de Florian and R.~Sassot,
410: Phys.\ Rev.\ D {\bf 62}, 094025 (2000).
411: \bibitem{deFlorian:2005mw}
412: D.~de Florian, G.~A.~Navarro and R.~Sassot,
413: \emph{Phys.\ Rev.\ D} \textbf{71}, 094018 (2005).
414: \bibitem{martin:2003}
415: A.D.Martin, R.G.Roberts, W.J.Stirling and R.S.Thorne, Eur.Phys.J.{\bf C28},
416: 455 (2003); J.Pumplin, D.R.Stump, J.Huston, H.L.Lai, P.Nadolsky and W.K.Tung,
417: J.High Energy Phys. 07 (2002) 012.
418: \bibitem{xiaodong} X. Jiang et al. Jefferson Lab experiment E04-113:
419: Semi-Inclusive Spin Asymmetries on the Nucleon Experiment (hep-ex/0412010).
420: \bibitem{kretzer} S. Kretzer,
421: \emph{Phys. Rev. D} \textbf{62} (2000) 054001.
422: \bibitem{inclusiva} R. Mertig, W. L. van Neerven, Z.Phys.{\bf C70}, 637,(1996);
423: W. Vogelsang, Phys.Rev.{\bf D54}, 2023, (1996).
424: \bibitem{newgr} M. Gluck, E. Reya, M. Stratmann and W. Vogelsang,
425: Phys.Rev.{\bf D53}, 4775 (1996).
426: \bibitem{NPB}D. de Florian, C.A. Garcia Canal,
427: and R. Sassot, Nucl. Phys. {\bf B470}, 195 (1996).
428: \bibitem{SMCi} SMC Collaboration, B. Adeva et al., Phys. Rev. {\bf D58}, 112001 (1998).
429: \bibitem{E143} E143 Collaboration, K. Abe et al, Phys. Rev. D {\bf 58}, 112003 (1998).
430: \bibitem{E155} E155 Collaboration, P. L. Anthony, et al., Phys. Lett. {\bf B463}, 339 (1999); G. S. Mitchell, Ph.D. Thesis University of Wisconsin-Madison, SLAC-Report-540 (1999)
431: \bibitem{E142} E142 Collaboration, P. L. Anthony, et al., Phys. Rev. D {\bf 54}, 6620 (1996).
432: \bibitem{E154} E154 Collaboration, K. Abe et al., Phys. Rev. Lett. {\bf 79}, 26 (1997); Phys. Lett. {\bf B405}, 180 (1997).
433: \bibitem{HALLA} HALL A Collaboration X. Zheng, et al., nucl-ex/0405006
434: \bibitem{SMC} B. Adeva et al., SMC Collab., Phys. Lett. {\bf B369}, 93 (1996), ibid. {\bf 420} 180 (1998).
435: \bibitem{HERMES} HERMES Collab., A. Airapetian, et al. hep-ex/0407032;
436: K. Ackerstaff et al,, Phys. Lett. {\bf B464}, 123 (1999).
437: \bibitem{COMPASS} A. Bressan, for the COMPASS Collaboration, hep-ex/0501040.
438: \bibitem{Stump:2001gu}
439: J.~Pumplin, D.~R.~Stump and W.~K.~Tung,
440: Phys.\ Rev.\ D {\bf 65}, 014011 (2002) \\
441: D.~Stump {\it et al.},
442: Phys.\ Rev.\ D {\bf 65}, 014012 (2002)
443: \bibitem{Ellis:2005cy}
444: J.~R.~Ellis and M.~Karliner,
445: %``Direct estimate of the gluon polarization in the nucleon,''
446: arXiv:hep-ph/0501115.
447: %%CITATION = HEP-PH 0501115;%%
448:
449: \end{thebibliography}
450: \end{document}
451:
452:
453: