hep-ex0312023/gc.tex
1: 
2: 
3: %
4: \section{Introduction}
5: \label{sec:gc_introduction}
6:  
7: %
8: %
9: %
10: %
11: %
12: %
13: 
14: The measurement of  gauge boson couplings and the
15: search for possible anomalous contributions due to the effects of new
16: physics beyond the Standard Model are among the principal physics
17: aims at \LEPII~\cite{gc_bib:LEP2YR}.
18: Combined preliminary measurements of triple gauge boson
19: couplings are presented here. Results from W-pair production are
20: combined in single and two-parameter fits, including updated results from
21: ALEPH, L3 and OPAL as well as an improved treatment of the main systematic
22: effect in our previous combination, the uncertainty in the $O(\alpha_{em})$
23: correction.     
24: An updated combination of quartic gauge coupling (QGC) results for the \ZZgg\
25: vertex is also presented, including data from ALEPH, L3 and OPAL.
26: The combination of QGCs associated with the \WWgg\ vertex, including the sign
27: convention as reported 
28: in~\cite{gc_bib:Montagna:2001ej,gc_bib:denner} and the reweighting based
29: on~\cite{gc_bib:Montagna:2001ej} is foreseen for our next report. 
30: The combination of neutral TGCs measured in ZZ production (f-couplings) has
31: been updated, including new results from L3 and OPAL.
32: The combinations for neutral TGCs accessible through ${\rm Z} \gamma$
33: production (h-couplings) reported in 2001 still remain valid ~\cite{gc_bib:budapest01}.
34: 
35: The W-pair production process, $\mathrm{e^+e^-\rightarrow\WW}$,
36: involves charged triple gauge boson vertices between the $\WW$ and
37: the Z or photon.  During \LEPII\ operation, about 10,000 W-pair
38: events were collected by each experiment.  
39: Single W ($\enw$) and single photon ($\nng$) production at LEP are also
40: sensitive to the $\WWg$ vertex. Results from these channels are also included
41: in the combination for some experiments; the individual references should be
42: consulted for details.
43: 
44: For the charged TGCs, Monte Carlo calculations
45: (RacoonWW~\cite{common_bib:racoonww} and
46: YFSWW~\cite{common_bib:yfsww}) incorporating an improved treatment of
47: $O(\alpha_{em})$ corrections to the WW production have become our standard by
48: now. The corrections affect the measurements of the charged
49: TGCs in W-pair production. 
50: Results, some of them preliminary, including these
51: $O(\alpha_{em})$ corrections have been submitted from all four LEP
52: collaborations ALEPH~\cite{gc_bib:ALEPH-cTGC}, DELPHI~\cite{gc_bib:DELPHI-cTGC}, L3~\cite{gc_bib:L3-cTGC} and OPAL~\cite{gc_bib:OPAL-cTGC3}. 
53: LEP combinations are made for the charged TGC measurements in single- and
54: two-parameter fits. 
55: 
56: At centre-of-mass energies exceeding twice the Z boson mass, pair
57: production of Z bosons is kinematically allowed. Here, one searches
58: for the possible existence of triple vertices involving only neutral
59: electroweak gauge bosons. Such vertices could also contribute to
60: Z$\gamma$ production.  In contrast to triple gauge boson vertices with
61: two charged gauge bosons, purely neutral gauge boson vertices do not
62: occur in the Standard Model of electroweak interactions.
63: 
64: Within the Standard Model, quartic electroweak gauge boson vertices
65: with at least two charged gauge bosons exist. In $\ee$ collisions at
66: \LepII\ centre-of-mass energies, the $\WWZg$ and $\WWgg$ vertices
67: contribute to $\WWg$ and $\nngg$ production in $s$-channel and
68: $t$-channel, respectively.  The effect of the Standard Model quartic
69: electroweak vertices is below the sensitivity of \LepII.  
70: Quartic gauge boson vertices with only neutral bosons, like the
71: \ZZgg\ vertex, do not exist in the Standard Model. However, anomalous QGCs
72: associated with this vertex are studied at LEP.
73: 
74: Anomalous quartic vertices are searched for in the production of
75: $\WWg$, $\nngg$ and $\Zgg$ final states. The couplings related to the \ZZgg\
76: and \WWgg\ vertices are assumed to be different~\cite{gc_bib:QGC-Belanger},
77: and are therefore treated separately.
78: In this report, we only combine the results for the anomalous couplings
79: associated with the \ZZgg\ vertex. The combination of the \WWgg\ vertex
80: couplings is foreseen for the near future.
81: 
82: 
83: \subsection{Charged Triple Gauge Boson Couplings}
84: \label{sec:gc_cTGCs}
85: 
86: The parametrisation of the charged triple gauge boson vertices is
87: described in References~\cite{gc_bib:GAEMERS,gc_bib:Hagiwara1987vm,
88: gc_bib:HAGIWARA,gc_bib:BILENKY,gc_bib:KUSS,
89: gc_bib:PAPADOPOULOSCP,gc_bib:LEP2YR}.  
90: The most general Lorentz invariant
91: Lagrangian which describes the triple gauge boson interaction has
92: fourteen independent complex couplings, seven describing the
93: WW$\gamma$ vertex and seven describing the WWZ vertex.  Assuming
94: electromagnetic gauge invariance as well as C and P conservation, the
95: number of independent TGCs reduces to five.  A common set is \{$\gz,
96: \kz, \kg, \lz$, $\lg$\} where $\gz = \kz = \kg = 1$ and $\lz = \lg =
97: 0$ in the \SM.  The parameters proposed in~\cite{gc_bib:LEP2YR} and used by
98: the LEP experiments are $\gz$, $\lg$ and $\kg$ with the gauge constraints:
99: \begin{eqnarray}
100: \kz & = & \gz - (\kg - 1) \twsq \,, \\
101: \lz & = & \lg \,,
102: \end{eqnarray}                               
103: where $\theta_W$ is the weak mixing angle.  The
104: couplings are considered as real, with the imaginary parts fixed to
105: zero. In contrast to previous LEP
106: combinations~\cite{gc_bib:moriond01,gc_bib:budapest01}, we are quoting
107: the measured coupling values themselves and not their deviation from the
108: Standard Model. 
109: 
110: Note that the photonic couplings $\lg$ and $\kg$ are related to the
111: magnetic and electric properties of the W-boson. One can write the
112: lowest order terms for a multipole expansion describing the W-$\gamma$
113: interaction as a function of $\lg$ and $\kg$. For the magnetic dipole
114: moment $\mu_{W}$ and the electric quadrupole moment $q_{W}$ one
115: obtains $e(1+\kappa_{\gamma}+\lambda_{\gamma})/2\MW$ and
116: $-e(\kappa_{\gamma}-\lambda_{\gamma})/\MW^2$, respectively.
117: %
118: %
119: %
120: %
121: %
122: %
123: %
124: 
125: The inclusion of $O(\alpha_{em})$ corrections in the Monte Carlo
126: calculations has a considerable effect on the charged TGC measurement. Both the
127: total cross-section and the differential distributions are affected. The  
128: cross-section is reduced by 1-2\% (depending on the energy). Amongst
129: the differential distributions, the effects are naturally more complex. The
130: polar W$^-$ production angle carries most of the information on the TGC
131: parameters; its 
132: shape is modified to be more forwardly peaked. In a fit to data, the
133: $O(\alpha_{em})$ effect
134: manifests itself as a negative shift of the obtained TGC values with a
135: magnitude of typically -0.015 for \lg\ and \gz\, and -0.04 for \kg.
136: 
137: \subsection{Neutral Triple Gauge Boson Couplings}
138: \label{sec:gc_nTGCs}
139: 
140: There are two classes of Lorentz invariant structures associated with
141: neutral TGC vertices which preserve $U(1)_{em}$ and Bose symmetry, as
142: described in~\cite{gc_bib:Hagiwara1987vm,gc_bib:Gounaris2000tb}.
143: 
144: The first class refers to anomalous Z$\gamma\gamma^*$ and $\rm Z\gamma
145: \rm Z^*$ couplings which are accessible at LEP in the process
146: $\mathrm{e^{+} e^{-}} \rightarrow {\rm Z} \gamma$. The parametrisation
147: contains eight couplings: $h_i^{V}$ with $i=1,...,4$ and $V=\gamma$,Z.
148: The superscript $\gamma$ refers to Z$\gamma\gamma^*$ couplings and
149: superscript Z refers to $\rm Z\gamma \rm Z^*$ couplings.  The photon
150: and the Z boson in the final state are considered as on-shell
151: particles, while the third boson at the vertex, the $s$-channel
152: internal propagator, is off shell.  The couplings $h_{1}^{V}$ and
153: $h_{2}^{V}$ are CP-odd while $h_{3}^{V}$ and $h_{4}^{V}$ are CP-even.
154: 
155: The second class refers to anomalous $\rm{ZZ}\gamma^*$ and
156: $\rm{ZZZ}^*$ couplings which are accessible at \LEPII\ in the process
157: $\mathrm{e^{+} e^{-}} \rightarrow$ ZZ.  This anomalous vertex is
158: parametrised in terms of four couplings: $f_{i}^{V}$ with $i=4,5$ and
159: $V=\gamma$,Z.  The superscript $\gamma$ refers to ZZ$\gamma^*$
160: couplings and the superscript Z refers to $\rm{ZZZ}^*$ couplings,
161: respectively.  Both Z bosons in the final state are assumed to be
162: on-shell, while the third boson at the triple vertex, the $s$-channel
163: internal propagator, is off-shell.
164: The couplings $f_{4}^{V}$ are CP-odd whereas $f_{5}^{V}$ are CP-even.
165: 
166: The $h_i^{V}$ and $f_{i}^{V}$ couplings are assumed to be real and they
167: vanish at tree level in the Standard Model.
168: 
169: 
170: \subsection{Quartic Gauge Boson Couplings}
171: \label{sec:gc_QGCs}
172: The couplings associated with the two QGC vertices \WWgg\ and \ZZgg\ are
173: assumed to be different, and are by convention treated as separate couplings
174: at LEP. In this report, we only combine QGCs related to the \ZZgg\ vertex.
175: The contribution of such anomalous quartic gauge boson couplings is described by two coupling parameters \acl\ and \azl,
176: which are zero in the Standard
177: Model~\cite{gc_bib:QGC-BelBou,4f_bib:eewwg}.  
178: Events from $\nngg$ and $\Zgg$ final states can originate from the \ZZgg\
179: vertex and are therefore used to study anomalous QGCs.
180: 
181: 
182: 
183: 
184: 
185: %
186: \section{Measurements}
187: \label{sec:gc_data}
188: 
189: The combined results presented here are obtained from charged and neutral
190: electroweak gauge boson coupling measurements, and from quartic gauge boson
191: couplings measurements as discussed above.  
192: The individual references should be consulted for details about the data
193: samples used. 
194: 
195: The charged TGC analyses of ALEPH, DELPHI, L3 and OPAL use data collected
196: at \LEPII\ up to centre-of-mass energies of 209~\GeV. These
197: analyses use different
198: channels, typically the semileptonic and fully hadronic W-pair
199: decays~\cite{gc_bib:ALEPH-cTGC,gc_bib:DELPHI-cTGC,
200:   gc_bib:L3-cTGC,gc_bib:OPAL-cTGC3}. 
201: The full data set is analysed by ALEPH, L3 and OPAL, whereas DELPHI presently
202: uses all data at 189 $\GeV$ and above. 
203: Anomalous
204: TGCs affect both the total production cross-section and the shape of
205: the differential cross-section as a function of the polar W$^-$
206: production angle.  The relative contributions of each helicity state
207: of the W bosons are also changed, which in turn affects the
208: distributions of their decay products.  The analyses presented by each
209: experiment make use of different combinations of each of these
210: quantities.  In general, however, all analyses use at least the
211: expected variations of the total production cross-section and the
212: W$^-$ production angle. Results from $\enw$ and $\nng$ production are
213: included by some 
214: experiments.  Single W production is particularly sensitive to \kg,
215: thus providing information complementary to that from W-pair
216: production.
217: 
218: The $h$-coupling analyses of ALEPH, DELPHI and L3 use data
219: collected up to centre-of-mass energies of 209~\GeV. The
220: OPAL measurements so far use the data at 189~\GeV.  The results of the
221: $f$-couplings are obtained from the whole data set above the
222: ZZ-production threshold by all of the experiments.  The experiments
223: already pre-combine different processes and final states for each of
224: the couplings.  For the neutral TGCs, the analyses use measurements of
225: the total cross sections of Z$\gamma$ and ZZ production and the
226: differential distributions: the $h_i^V$
227: couplings~\cite{gc_bib:ALEPH-nTGC,gc_bib:DELPHI-nTGC,gc_bib:L3-hTGC,
228:   gc_bib:OPAL-hTGC} and the $f_i^V$
229: couplings~\cite{gc_bib:ALEPH-nTGC,gc_bib:DELPHI-nTGC,gc_bib:L3-fTGC,
230:   gc_bib:OPAL-fTGC} are determined.
231: 
232: The combination of quartic gauge boson couplings associated with
233: the \ZZgg\ vertex is at present based on analyses of
234: ALEPH~\cite{gc_bib:ALEPH-QGC}, L3~\cite{gc_bib:L3-QGC} 
235: and OPAL~\cite{gc_bib:OPAL-QGC}.
236: The L3 analysis uses data from the \qqgg\
237: final state all at centre-of-mass energies above the Z resonance, from 130
238: GeV to 207 GeV. 
239: Both ALEPH and OPAL analyse the $\nngg$ final state, with ALEPH using data
240: from centre-of-mass energies ranging from 183 GeV to 209 GeV, and OPAL from
241: 189 GeV to 209 GeV.  
242: 
243: %
244: \section{Combination Procedure}
245: \label{sec:gc_combination}
246: 
247: The combination is based on the individual likelihood functions from the four
248: LEP experiments.
249: Each experiment provides the negative log likelihood, $\LL$, as a
250: function of the coupling parameters to be combined.  
251: The single-parameter analyses are performed fixing 
252: all other parameters to their Standard Model values.  The
253: two-parameter analyses are performed setting the remaining
254: parameters to their Standard Model values. For the charged TGCs, the
255: gauge constraints listed in Section~\ref{sec:gc_cTGCs} are always
256: enforced.
257: 
258: The $\LL$ functions from each experiment include statistical as well
259: as those systematic uncertainties which are considered as
260: uncorrelated between experiments.  For both single- and
261: multi-parameter combinations, the individual $\LL$ functions are
262: combined.  It is necessary to use the $\LL$ functions directly in the
263: combination, since in some cases they are not parabolic, and hence it is not
264: possible to properly combine the results by simply taking weighted
265: averages of the measurements.
266: 
267: The main contributions to the systematic uncertainties that are
268: uncorrelated between experiments arise from detector effects,
269: background in the selected signal samples, limited Monte Carlo
270: statistics and the fitting method.  Their importance varies for each
271: experiment and the individual references should be consulted for
272: details.
273: 
274: In the neutral TGC sector, the systematic uncertainties arising from the
275: theoretical cross 
276: section prediction in Z$\gamma$-production ($\simeq 1\%$ in the
277: $\qq\gamma$- and $\simeq 2\%$ in the $\nng$ channel) are treated as
278: correlated.
279: For ZZ production, the uncertainty on the theoretical cross section
280: prediction is small compared to the statistical accuracy and therefore
281: is neglected.  Smaller sources of correlated systematic uncertainties,
282: such as those arising from the LEP beam energy, are for simplicity
283: treated as uncorrelated.
284: 
285: The combination procedure for neutral TGCs, where the relative
286: systematic uncertainties are small, is unchanged with respect to the
287: previous LEP combinations of electroweak gauge boson
288: couplings~\cite{gc_bib:moriond01,gc_bib:budapest01}. 
289: The correlated systematic uncertainties in the $h$-coupling analyses
290: are taken into account by scaling the combined log-likelihood
291: functions by the squared ratio of the sum of statistical and
292: uncorrelated systematic uncertainty over the total uncertainty
293: including all correlated uncertainties.  For the general case of
294: non-Gaussian probability density functions, this treatment of the
295: correlated errors is only an approximation; it also neglects
296: correlations in the systematic uncertainties between the parameters in
297: multi-parameter analyses.
298: 
299: In the charged TGC sector, systematic uncertainties considered
300: correlated between the experiments are the theoretical cross
301: section prediction ($0.5\%$ for W-pair production and $5\%$ for single W
302: production), hadronisation effects, the final
303: state interactions, namely Bose-Einstein correlations and colour reconnection,
304: and the uncertainty in the radiative corrections themselves. 
305: The latter was the dominant systematic error in our previous combination,
306: where we used a conservative estimate, the full effect from applying the
307: $O(\alpha_{em})$ corrections. 
308: New preliminary analyses on the subject are now available from several LEP
309: experiments~\cite{gc_bib:ALEPH-cTGC}, based on comparisons 
310: of fully simulated events using two different leading-pole approximation
311: schemes (LPA-A and LPA-B)~\cite{gc_bib:LPA_A-B}.
312: In addition, the availability of comparisons of both generators
313: incorporating $O(\alpha_{em})$ corrections (RacoonWW and
314: YFSWW~\cite{common_bib:racoonww,common_bib:yfsww}) makes 
315: it now possible to perform a more realistic estimation of this effect. 
316: In general, the TGC shift
317: measured in the comparison of the two generators is found to be larger than
318: the effect from the different LPA schemes.
319: This improved
320: estimation, whilst still being conservative, reduces the systematic
321: uncertainty from $O(\alpha_{em})$ corrections by about a
322: third for $\gz$ and $\lg$ and roughly halves it for $\kg$, compared to the
323: full $O(\alpha_{em})$ correction applied previously. The application of this
324: reduced systematic error renders the charged TGC measurements statistics
325: dominated. 
326: 
327: In case of the charged TGCs, the systematic uncertainties considered
328: correlated between the experiments amount to 58\% of the combined
329: statistical and uncorrelated uncertainties for $\lg$ and $\gz$, while for
330: $\kg$ it is 68\%.  This means that the measurements of $\lg$, $\gz$ and $\kg$
331: are now clearly limited by statistics.
332: An improved combination
333: procedure~\cite{gc_bib:Alcaraz} is used for the charged TGCs.  
334: This procedure allows the combination of statistical and correlated
335: systematic uncertainties, independently of the analysis method chosen by
336: the individual experiments. 
337: 
338: The combination of charged TGCs uses the likelihood curves and
339: correlated systematic errors submitted by each of the four experiments. 
340: The procedure is based on the introduction of an additional free parameter to
341: take into account the systematic uncertainties, which are treated as shifts on
342: the fitted TGC value, and are assumed to have a Gaussian distribution. 
343: A simultaneous minimisation of both parameters (TGC
344: and systematic error) is performed to the log-likelihood function. 
345: 
346: In detail, the combination proceeds in the following way: the set of
347: measurements from the LEP experiments
348: ALEPH, DELPHI, OPAL and L3 is given with statistical plus uncorrelated
349: systematic uncertainties in terms of likelihood curves:  
350: $-\log{\mathcal L}^A_{stat}(x)$,
351: $-\log{\mathcal L}^D_{stat}(x)$ 
352: $-\log{\mathcal L}^L_{stat}(x)$ 
353: and $-\log{\mathcal L}^O_{stat}(x)$, 
354: respectively, where $x$ is the coupling parameter in question. 
355: Also given are the shifts for each of the five totally correlated sources
356: of uncertainty mentioned above; each source $S$ is 
357: leading to systematic errors $\sigma^S_A$, $\sigma^S_D$, $\sigma^S_L$ and
358: $\sigma^S_O$.
359: 
360: Additional parameters $\Delta^S$ are included in order to take into 
361: account a Gaussian distribution for each of the systematic uncertainties.
362: The procedure then consists in minimising the function:
363: \noindent
364: \begin{eqnarray}
365: -\log {\mathcal L}_{total} = 
366: \sum_{E=A,D,L,O} \log {\mathcal L}^E_{stat} 
367: (x-\sum_{S=DPA,\sigma_{WW},HAD,BE,CR}(\sigma^S_E \Delta^S))
368:  + \sum_{S} {\displaystyle \frac{(\Delta^S)^{2}}{2}} \\ \nonumber
369: \end{eqnarray}
370: 
371: \noindent
372: where $x$ and $\Delta_S$ are the free parameters, and the sums run over the
373: four experiments and the five systematic errors.
374: The resulting uncertainty on $x$ will take into account all sources 
375: of uncertainty, yielding a measurement of the coupling with the error
376: representing statistical and systematic sources.
377: The projection of the minima of the log-likelihood as a function of $x$ 
378: gives the combined log-likelihood curve including statistical and
379: systematic uncertainties. 
380: The advantage over the scaling method used previously is that it
381: treats systematic uncertainties that are correlated between the experiments
382: correctly, while not forcing the averaging of these systematic
383: uncertainties into one global LEP systematics scaling factor. In other words,
384: the (statistical) precision of each experiment now gets reduced by its own
385: correlated systematic errors, instead of an averaged LEP systematic
386: error. 
387: The method has been cross-checked against the scaling method, and was found
388: to give comparable results. 
389: The inclusion of
390: the systematic uncertainties lead to small differences as expected by the
391: improved treatment of correlated systematic errors, a similar behaviour as
392: seen in Monte Carlo comparisons of these two combinations methods
393: ~\cite{gc_bib:renaud}. Furthermore, it was shown that the minimisation-based
394: combination method used for the charged TGCs agrees with the method
395: based on optimal observables, where systematic effects are included directly
396: in the mean values of the optimal observables (see~\cite{gc_bib:renaud}), 
397: for any realistic ratio of statistical and systematic uncertainties. 
398: Further details on the improved combination method can be found
399: in~\cite{gc_bib:Alcaraz}. 
400: 
401: In the combination of the QGCs, the influence of correlated systematic
402: uncertainties is considered negligible compared to the statistical error,
403: arising from the small number of selected events. Therefore, the QGCs are
404: combined by adding the log-likelihood curves from the single experiments. 
405: 
406: For all single- and multi-parameter results quoted in numerical form,
407: the one standard deviation uncertainties (68\% confidence level) are
408: obtained by taking the coupling values for which $\Delta\LL=+0.5$
409: above the minimum.  The 95\% confidence level (C.L.)  limits are given
410: by the coupling values for which $\Delta\LL=+1.92$ above the minimum.
411: Note that in the case of the neutral TGCs, double minima structures
412: appear in the negative log-likelihood curves.  For multi-parameter
413: analyses, the two dimensional 68\%~C.L.~contour curves for any pair of
414: couplings are obtained by requiring $\Delta\LL=+1.15$, while for the
415: 95\% C.L.~contour curves $\Delta\LL=+3.0$ is required.  Since the
416: results on the different parameters and parameter sets are obtained
417: from the same data sets, they cannot be combined.
418: 
419: \section{Results}
420: 
421: We present results from the four LEP experiments on the various
422: electroweak gauge boson couplings, and their combination.
423: %
424: %
425: The charged TGC combination has been updated with the inclusion of recent
426: results from ALEPH, L3 and OPAL.
427: The neutral TGC results include an update of the $f_i^V$ combinations,
428: whilst the $h_i^V$ combinations remain unchanged since
429: our last note~\cite{gc_bib:budapest01}.  The results
430: quoted for each individual experiment are calculated using the methods
431: described in Section~\ref{sec:gc_combination}.  Therefore they may differ
432: slightly from those reported in the individual references, as the experiments
433: in general use other methods to combine the data from different channels, and
434: to include systematic uncertainties. 
435: In particular for the charged couplings, experiments using a
436: combination method based on optimal observables (ALEPH, OPAL) 
437: obtain results with small differences
438: compared to the values given by our combination technique. These small
439: differences have been studied in Monte Carlo tests and are well
440: understood~\cite{gc_bib:renaud}.  
441: For the $h$-coupling result from OPAL and DELPHI, a slightly
442: modified estimate of the systematic uncertainty due to the theoretical
443: cross section prediction is responsible for slightly different limits
444: compared to the published results.  
445: 
446: 
447: \subsection{Charged Triple Gauge Boson Couplings}
448: 
449: The individual analyses and results of the experiments for the charged couplings are described in~\cite{gc_bib:ALEPH-cTGC,gc_bib:DELPHI-cTGC,
450: gc_bib:L3-cTGC,gc_bib:OPAL-cTGC3}.
451: 
452: \subsubsection*{Single-Parameter Analyses}
453: The results of single-parameter fits from each experiment are shown in
454: Table~\ref{tab:cTGC-1-ADLO}, where the errors include both statistical
455: and systematic effects. The individual $\LL$ curves and their sum are shown in
456: Figure~\ref{fig:cTGC-1}.  The results of the combination are given in
457: Table~\ref{tab:cTGC-1-LEP}. A list of the systematic errors treated as fully
458: correlated between the LEP experiments, and their shift on the combined fit
459: result are given in Table ~\ref{tab:cTGC-syst}.
460: 
461: \subsubsection*{Two-Parameter Analyses}
462: Contours at 68\% and 95\% confidence level for the combined two-parameter
463: fits are shown in Figure~\ref{fig:cTGC-2}. The numerical results of the
464: combination are given in Table~\ref{tab:cTGC-2D-LEP}. The errors include both statistical and systematic effects. 
465: 
466: \begin{table}[htbp]
467: \begin{center}
468: \renewcommand{\arraystretch}{1.3}
469: \begin{tabular}{|l||r|r|r|r|} 
470: \hline
471: Parameter  & ALEPH   & DELPHI  &  L3   & OPAL  \\
472: \hline
473: \hline
474: \gz       & $1.026\apm{0.034}{0.033}$ & $1.002\apm{0.038}{0.040}$ 
475:            & $0.928\apm{0.042}{0.041}$ & $0.985\apm{0.035}{0.034}$ \\ 
476: \hline
477: \kg       & $1.022\apm{0.073}{0.072}$ & $0.955\apm{0.090}{0.086}$  
478:            & $0.922\apm{0.071}{0.069}$ & $0.929\apm{0.085}{0.081}$ \\  
479: \hline
480: \lg        & $0.012\apm{0.033}{0.032}$ & $0.014\apm{0.044}{0.042}$  
481:            & $-0.058\apm{0.047}{0.044}$ & $-0.063\apm{0.036}{0.036}$ \\
482: \hline
483: \end{tabular}
484: \caption[]{The measured central values and one standard deviation errors
485:   obtained by the four LEP experiments.  In each case the parameter
486:   listed is varied while the remaining two are fixed to their Standard Model
487:   values. Both
488:   statistical and systematic errors are included. The values given here
489:   differ slightly from the ones quoted in the individual contributions from
490:   the four LEP experiments, as a different combination method is used. See
491:   text in section \ref{sec:gc_combination} for details. }
492: \label{tab:cTGC-1-ADLO}
493: \end{center}
494: \end{table}
495: 
496: \begin{table}[htbp]
497: \begin{center}
498: \renewcommand{\arraystretch}{1.3}
499: \begin{tabular}{|l||r|c|} 
500: \hline
501: Parameter  & 68\% C.L.   & 95\% C.L.      \\
502: \hline
503: \hline
504: $\gz$     & $0.991\apm{0.022}{0.021} $  & [$0.949,~~1.034$]  \\ 
505: \hline
506: $\kg$     & $0.984\apm{0.042}{0.047}$  & [$0.895,~~1.069$]  \\ 
507: \hline
508: $\lg$     & $-0.016\apm{0.021}{0.023}$  & [$-0.059,~~0.026$]  \\ 
509: \hline
510: \end{tabular}
511: \caption[]{ The combined 68\% C.L. errors and 95\% C.L. intervals
512:   obtained combining the results from the four LEP experiments.  In
513:   each case the parameter listed is varied while the other two are
514:   fixed to their Standard Model values.  Both statistical and systematic
515:   errors are included.  }
516:  \label{tab:cTGC-1-LEP}
517: \end{center}
518: \end{table}
519: 
520: \begin{table}[htbp]
521: \begin{center}
522: \renewcommand{\arraystretch}{1.3}
523: \begin{tabular}{|l||r|r|r|} 
524: \hline
525: Source  & \gz  & \lg   & \kg  \\
526: \hline
527: \hline
528: $O(\alpha_{em})$ correction  & 0.010 & 0.010  & 0.020 \\ 
529: $\sigma_{WW}$ prediction   & 0.003  & 0.005 & 0.014 \\ 
530: Hadronisation   & 0.004 & 0.002 & 0.004 \\ 
531: Bose-Einstein Correlation   & 0.005  & 0.004  & 0.009 \\ 
532: Colour Reconnection   & 0.005 & 0.004 & 0.010 \\
533: $\sigma_{single W}$ prediction  & - & - &  0.011 \\ 
534: \hline
535: \end{tabular}
536: \caption[]{The systematic uncertainties considered correlated between the LEP
537:   experiments in the charged TGC combination and their effect on the combined
538:   fit results.}
539: \label{tab:cTGC-syst}
540: \end{center}
541: \end{table}
542: 
543: \begin{table}[htbp]
544: \begin{center}
545: \begin{tabular}{|l||r|r|rr|} \hline
546: Parameter   & 68\% C.L.  & 95\% C.L.   & \multicolumn{2}{|c|}{Correlations}  \\
547: \hline \hline
548: \gz       & $1.004\apm{0.024}{0.025}$ & [$+0.954,~~+1.050$]
549:   & 1.00 & +0.11 \\  
550: \kg       & $0.984\apm{0.049}{0.049}$ & [$+0.894,~~+1.084$]  
551:  & +0.11 & 1.00 \\ 
552: \hline
553: \gz       & $1.024\apm{0.029}{0.029}$ & [$+0.966,~~+1.081$] & 1.00
554:   & -0.40  \\  
555: \lg       & $-0.036\apm{0.029}{0.029}$ & [$-0.093,~~+0.022$] 
556:  & -0.40  & 1.00 \\ 
557: \hline
558: \kg        & $1.026\apm{0.048}{0.051}$ & [$+0.928,~~+1.127$]
559:  & 1.00  & +0.21 \\ 
560: \lg        & $-0.024\apm{0.025}{0.021}$ & [$-0.068,~~+0.023$] 
561:   & +0.21 & 1.00 \\ 
562: \hline
563: \end{tabular}
564: \end{center}
565: \caption{ The measured central values, one standard deviation errors and
566:   limits at 95\% confidence level, 
567:     obtained by combining the four LEP experiments for the 
568:     two-parameter fits of the charged TGC parameters.
569:     Since the shape of the log-likelihood
570:   is not parabolic, there is some ambiguity in the definition of the
571:   correlation coefficients and the values quoted here are approximate.
572:     The listed parameters are varied while the
573:     remaining one is fixed to its Standard Model value.
574:     Both statistical and systematic errors are included.
575:   }
576: \label{tab:cTGC-2D-LEP}
577: \end{table}
578: 
579: \clearpage
580: 
581: \begin{figure}[htbp]
582: \begin{center}
583: \includegraphics[width=\linewidth]{gc_ctgc1d_aachen03.eps}
584: \caption[]{
585:   The $\LL$ curves of the four experiments (thin lines) and the LEP
586:   combined curve (thick line) for the three charged TGCs $\gz$,
587:   $\kg$ and $\lg$.  In each case, the minimal value is subtracted.  }
588: \label{fig:cTGC-1}
589: \end{center}
590: \end{figure}
591: 
592: \begin{figure}[htbp]
593: \begin{center}
594: \includegraphics[width=\linewidth]{gc_ctgc2d_aachen03.eps}
595: \caption[]{
596: The 68\% and 95\% confidence level contours for the three two-parameter fits
597: to the charged TGCs $\gz$-$\lg$, $\gz$-$\kg$ and $\lg$-$\kg$. 
598: The fitted coupling
599: value is indicated with a cross; the Standard Model value for each fit is in
600: the centre of the grid. The contours include the contribution from systematic
601: uncertainties.}
602:  \label{fig:cTGC-2}
603: \end{center}
604: \end{figure}
605: 
606: \clearpage
607: 
608: \subsection{Neutral Triple Gauge Boson Couplings in Z\boldmath$\gamma$ 
609: Production}
610: 
611: The individual analyses and results of the experiments for the $h$-couplings 
612: are described in~\cite{gc_bib:ALEPH-nTGC,gc_bib:DELPHI-nTGC,
613: gc_bib:L3-hTGC,gc_bib:OPAL-hTGC}.
614: 
615: \subsubsection*{Single-Parameter Analyses}
616: The results for each experiment are shown in
617: Table~\ref{tab:gc_hTGC-1-ADLO}, where the errors include both
618: statistical and systematic uncertainties.  The individual $\LL$ curves
619: and their sum are shown in Figures~\ref{fig:gc_hgTGC-1}
620: and~\ref{fig:gc_hzTGC-1}.  The results of the combination are given in
621: Table~\ref{tab:gc_hTGC-1-LEP}.  From
622: Figures~\ref{fig:gc_hgTGC-1} and \ref{fig:gc_hzTGC-1} it is clear that the
623: sensitivity of the L3 analysis~\cite{gc_bib:L3-hTGC} is the highest
624: amongst the LEP experiments. This is partially due to the use of a larger
625: phase space region, which increases the statistics by about a factor two,
626: and partially due to additional information from using an optimal-observable
627: technique.
628: 
629: \subsubsection*{Two-Parameter Analyses}
630: 
631: The results for each experiment are shown in
632: Table~\ref{tab:gc_hTGC-2-ADLO}, where the errors include both
633: statistical and systematic uncertainties.  The 68\% C.L. and 95\% C.L.
634: contour curves resulting from the combinations of the two-dimensional
635: likelihood curves are shown in Figure~\ref{fig:gc_hTGC-2}.  The LEP
636: average values are given in Table~\ref{tab:gc_hTGC-2-LEP}.
637: 
638: %
639: 
640: \begin{table}[htbp]
641: \begin{center}
642: \renewcommand{\arraystretch}{1.3}
643: \begin{tabular}{|l||r|r|r|r|} 
644: \hline
645: Parameter  & ALEPH & DELPHI  &  L3  & OPAL \\
646: \hline
647: \hline
648: $h_1^\gamma$ & [$-0.14,~~+0.14$]  & [$-0.15,~~+0.15$]   & [$-0.06,~~+0.06$]   & [$-0.13,~~+0.13$] \\
649: \hline                             
650: $h_2^\gamma$ & [$-0.07,~~+0.07$]  & [$-0.09,~~+0.09$]   & [$-0.053,~~+0.024$] & [$-0.089,~~+0.089$] \\
651: \hline                             
652: $h_3^\gamma$ & [$-0.069,~~+0.037$]& [$-0.047,~~+0.047$] & [$-0.062,~~-0.014$] & [$-0.16,~~+0.00$] \\
653: \hline                             
654: $h_4^\gamma$ & [$-0.020,~~+0.045$]& [$-0.032,~~+0.030$] & [$-0.004,~~+0.045$] & [$+0.01,~~+0.13$] \\
655: \hline                             
656: $h_1^Z$      & [$-0.23,~~+0.23$]  & [$-0.24,~~+0.25$]   & [$-0.17,~~+0.16$]   & [$-0.22,~~+0.22$] \\
657: \hline                             
658: $h_2^Z$      & [$-0.12,~~+0.12$]  & [$-0.14,~~+0.14$]   & [$-0.10,~~+0.09$]   & [$-0.15,~~+0.15$] \\
659: \hline                             
660: $h_3^Z$      & [$-0.28,~~+0.19$]  & [$-0.32,~~+0.18$]   & [$-0.23,~~+0.11$]   & [$-0.29,~~+0.14$] \\
661: \hline                             
662: $h_4^Z$      & [$-0.10,~~+0.15$]  & [$-0.12,~~+0.18$]   & [$-0.08,~~+0.16$]   & [$-0.09,~~+0.19$] \\
663: \hline
664: \end{tabular}
665: \caption[]{The 95\% C.L. intervals ($\Delta\LL=1.92$) measured by
666:   the ALEPH, DELPHI, L3 and OPAL.  In each case the parameter listed is varied
667:   while the remaining ones are fixed to their Standard Model values.
668:   Both statistical and systematic uncertainties are included.  }
669: \label{tab:gc_hTGC-1-ADLO}
670: \end{center}
671: \end{table}
672: 
673: \begin{table}[htbp]
674: \begin{center}
675: \renewcommand{\arraystretch}{1.3}
676: \begin{tabular}{|l||c|} 
677: \hline
678: Parameter     & 95\% C.L.      \\
679: \hline
680: \hline
681: $h_1^\gamma$  & [$-0.056,~~+0.055$]  \\ 
682: \hline
683: $h_2^\gamma$  & [$-0.045,~~+0.025$]  \\ 
684: \hline
685: $h_3^\gamma$  & [$-0.049,~~-0.008$]  \\ 
686: \hline
687: $h_4^\gamma$  & [$-0.002,~~+0.034$]  \\ 
688: \hline
689: $h_1^Z$       & [$-0.13,~~+0.13$]  \\ 
690: \hline
691: $h_2^Z$       & [$-0.078,~~+0.071$]  \\ 
692: \hline
693: $h_3^Z$       & [$-0.20,~~+0.07$]  \\ 
694: \hline
695: $h_4^Z$       & [$-0.05,~~+0.12$]  \\ 
696: \hline
697: \end{tabular}
698: \caption[]{ The 95\% C.L. intervals ($\Delta\LL=1.92$) obtained
699:   combining the results from the four experiments.  In each case the
700:   parameter listed is varied while the remaining ones are fixed to
701:   their Standard Model values.  Both statistical and systematic
702:   uncertainties are included.  }
703:  \label{tab:gc_hTGC-1-LEP}
704: \end{center}
705: \end{table}
706: 
707: \begin{table}[htbp]
708: \begin{center}
709: \renewcommand{\arraystretch}{1.3}
710: \begin{tabular}{|l||r|r|r|} 
711: \hline
712: Parameter  & ALEPH & DELPHI  &  L3  \\
713: \hline
714: \hline
715: $h_1^\gamma$ & [$-0.32,~~+0.32$] & [$-0.28,~~+0.28$] & [$-0.17,~~+0.04$] \\
716: $h_2^\gamma$ & [$-0.18,~~+0.18$] & [$-0.17,~~+0.18$] & [$-0.12,~~+0.02$] \\
717: \hline                                               
718: $h_3^\gamma$ & [$-0.17,~~+0.38$] & [$-0.48,~~+0.20$] & [$-0.09,~~+0.13$] \\
719: $h_4^\gamma$ & [$-0.08,~~+0.29$] & [$-0.08,~~+0.15$] & [$-0.04,~~+0.11$] \\
720: \hline                                               
721: $h_1^Z$      & [$-0.54,~~+0.54$] & [$-0.45,~~+0.46$] & [$-0.48,~~+0.33$] \\
722: $h_2^Z$      & [$-0.29,~~+0.30$] & [$-0.29,~~+0.29$] & [$-0.30,~~+0.22$] \\
723: \hline
724: $h_3^Z$      & [$-0.58,~~+0.52$] & [$-0.57,~~+0.38$] & [$-0.43,~~+0.39$] \\
725: $h_4^Z$      & [$-0.29,~~+0.31$] & [$-0.31,~~+0.28$] & [$-0.23,~~+0.28$] \\
726: \hline
727: \end{tabular}
728: \caption[]{The 95\% C.L. intervals ($\Delta\LL=1.92$) measured  by
729:   ALEPH, DELPHI and L3.  In each case the two parameters listed are varied
730:   while the remaining ones are fixed to their Standard Model values.
731:   Both statistical and systematic uncertainties are included.  }
732: \label{tab:gc_hTGC-2-ADLO}
733: \end{center}
734: \end{table}
735: 
736: \begin{table}[htbp]
737: \begin{center}
738: \renewcommand{\arraystretch}{1.3}
739: \begin{tabular}{|l||c|rr|} 
740: \hline
741: Parameter  & 95\% C.L. & \multicolumn{2}{|c|}{Correlations} \\
742: \hline
743: \hline
744: $h_1^\gamma$  & [$-0.16,~~+0.05$]    & $ 1.00$ & $+0.79$ \\ 
745: $h_2^\gamma$  & [$-0.11,~~+0.02$]    & $+0.79$ & $ 1.00$ \\ 
746: \hline
747: $h_3^\gamma$  & [$-0.08,~~+0.14$]    & $ 1.00$ & $+0.97$ \\ 
748: $h_4^\gamma$  & [$-0.04,~~+0.11$]    & $+0.97$ & $ 1.00$ \\ 
749: \hline
750: $h_1^Z$       & [$-0.35,~~+0.28$]    & $ 1.00$ & $+0.77$ \\ 
751: $h_2^Z$       & [$-0.21,~~+0.17$]    & $+0.77$ & $ 1.00$ \\ 
752: \hline
753: $h_3^Z$       & [$-0.37,~~+0.29$]    & $ 1.00$ & $+0.76$ \\ 
754: $h_4^Z$       & [$-0.19,~~+0.21$]    & $+0.76$ & $ 1.00$ \\ 
755: \hline
756: \end{tabular}
757: \caption[]{ The 95\% C.L. intervals ($\Delta\LL=1.92$) obtained
758:   combining the results from ALEPH, DELPHI and L3.  In each case the two
759:   parameters listed are varied while the remaining ones are fixed to
760:   their Standard Model values.  Both statistical and systematic
761:   uncertainties are included.  Since the shape of the log-likelihood
762:   is not parabolic, there is some ambiguity in the definition of the
763:   correlation coefficients and the values quoted here are approximate.
764:   }
765:  \label{tab:gc_hTGC-2-LEP}
766: \end{center}
767: \end{table}
768: 
769: \clearpage
770: 
771: \begin{figure}[p]
772: \begin{center}
773: \includegraphics[width=\linewidth]{gc_bp01_hig.eps}
774: \caption[]{
775:   The $\LL$ curves of the four experiments, and the LEP combined curve
776:   for the four neutral TGCs $h_i^\gamma,~i=1,2,3,4$. In each case, the
777:   minimal value is subtracted.  }
778: \label{fig:gc_hgTGC-1}
779: \end{center}
780: \end{figure}
781: 
782: \begin{figure}[p]
783: \begin{center}
784: \includegraphics[width=\linewidth]{gc_bp01_hiz.eps}
785: \caption[]{
786:   The $\LL$ curves of the four experiments, and the LEP combined curve
787:   for the four neutral TGCs $h_i^Z,~i=1,2,3,4$.  In each case, the
788:   minimal value is subtracted.  }
789: \label{fig:gc_hzTGC-1}
790: \end{center}
791: \end{figure}
792: 
793: \begin{figure}[p]
794: \begin{center}
795: \includegraphics[width=0.49\linewidth]{gc_bp01_h1gh2g.eps}\hfill
796: \includegraphics[width=0.49\linewidth]{gc_bp01_h3gh4g.eps}\\
797: \includegraphics[width=0.49\linewidth]{gc_bp01_h1zh2z.eps}\hfill
798: \includegraphics[width=0.49\linewidth]{gc_bp01_h3zh4z.eps}
799: \caption[]{
800:   Contour curves of 68\% C.L. and 95\% C.L. in the planes
801:   $(h_1^\gamma,h_2^\gamma)$, $(h_3^\gamma,h_4^\gamma)$,
802:   $(h_1^Z,h_2^Z)$ and $(h_3^Z,h_4^Z)$ showing the LEP combined result.
803:   }
804: \label{fig:gc_hTGC-2}
805: \end{center}
806: \end{figure}
807: 
808: \clearpage
809: 
810: \subsection{Neutral Triple Gauge Boson Couplings in ZZ Production}
811: 
812: The individual analyses and results of the experiments for the $f$-couplings 
813: are described in~\cite{gc_bib:ALEPH-nTGC,gc_bib:DELPHI-nTGC,gc_bib:L3-fTGC,gc_bib:OPAL-fTGC}.
814: 
815: \subsubsection*{Single-Parameter Analyses}
816: 
817: The results for each experiment are shown in
818: Table~\ref{tab:gc_fTGC-1-ADLO}, where the errors include both statistical
819: and systematic uncertainties.  The individual $\LL$ curves and their sum are
820: shown in Figure~\ref{fig:gc_fTGC-1}.  The results of the combination are
821: given in Table~\ref{tab:gc_fTGC-1-LEP}.
822: 
823: \subsubsection*{Two-Parameter Analyses}
824: 
825: The results from each experiment are shown in
826: Table~\ref{tab:gc_fTGC-2-ADLO}, where the errors include both
827: statistical and systematic uncertainties.  The 68\% C.L. and 95\% C.L.
828: contour curves resulting from the combinations of the two-dimensional
829: likelihood curves are shown in Figure~\ref{fig:gc_fTGC-2}.  The LEP
830: average values are given in Table~\ref{tab:gc_fTGC-2-LEP}.
831: 
832: \begin{table}[htbp]
833: \begin{center}
834: \renewcommand{\arraystretch}{1.3}
835: \begin{tabular}{|l||r|r|r|r|} 
836: \hline
837: Parameter  & ALEPH & DELPHI  &  L3   & OPAL  \\
838: \hline
839: \hline
840: $f_4^\gamma$ & [$-0.26,~~+0.26$] & [$-0.26,~~+0.28$] & [$-0.28,~~+0.28$] & [$-0.32,~~+0.33$] \\ 
841: \hline
842: $f_4^Z$      & [$-0.44,~~+0.43$] & [$-0.49,~~+0.42$] & [$-0.48,~~+0.46$] & [$-0.45,~~+0.58$] \\ 
843: \hline
844: $f_5^\gamma$ & [$-0.54,~~+0.56$] & [$-0.48,~~+0.61$] & [$-0.39,~~+0.47$] & [$-0.71,~~+0.59$] \\ 
845: \hline
846: $f_5^Z$      & [$-0.73,~~+0.83$] & [$-0.42,~~+0.69$] & [$-0.35,~~+1.03$] & [$-0.94,~~+0.25$] \\ 
847: \hline
848: \end{tabular}
849: \caption[]{The 95\% C.L. intervals ($\Delta\LL=1.92$) measured by
850:   ALEPH, DELPHI, L3 and OPAL.  In each case the parameter listed is varied
851:   while the remaining ones are fixed to their Standard Model values.
852:   Both statistical and systematic uncertainties are included.  }
853: \label{tab:gc_fTGC-1-ADLO}
854: \end{center}
855: \end{table}
856: 
857: \begin{table}[htbp]
858: \begin{center}
859: \renewcommand{\arraystretch}{1.3}
860: \begin{tabular}{|l||c|} 
861: \hline
862: Parameter     & 95\% C.L.     \\
863: \hline
864: \hline
865: $f_4^\gamma$  & [$-0.17,~~+0.19$]  \\ 
866: \hline
867: $f_4^Z$       & [$-0.30,~~+0.30$]  \\ 
868: \hline
869: $f_5^\gamma$  & [$-0.32,~~+0.36$]  \\ 
870: \hline
871: $f_5^Z$       & [$-0.34,~~+0.38$]  \\ 
872: \hline
873: \end{tabular}
874: \caption[]{ The 95\% C.L. intervals ($\Delta\LL=1.92$) obtained
875:   combining the results from all four experiments.  In each case the
876:   parameter listed is varied while the remaining ones are fixed to
877:   their Standard Model values.  Both statistical and systematic
878:   uncertainties are included.  }
879:  \label{tab:gc_fTGC-1-LEP}
880: \end{center}
881: \end{table}
882: 
883: \begin{table}[htbp]
884: \begin{center}
885: \renewcommand{\arraystretch}{1.3}
886: \begin{tabular}{|l||r|r|r|r|} 
887: \hline
888: Parameter  & ALEPH & DELPHI  &  L3   & OPAL  \\
889: 
890: \hline
891: \hline
892: $f_4^\gamma$ & [$-0.26,~~+0.26$] & [$-0.26,~~+0.28$]& [$-0.28,~~+0.28$] & [$-0.32,~~+0.33$] \\ 
893: $f_4^Z$      & [$-0.44,~~+0.43$] & [$-0.49,~~+0.42$]& [$-0.48,~~+0.46$] & [$-0.47,~~+0.58$]  \\ 
894: \hline                                              
895: $f_5^\gamma$ & [$-0.52,~~+0.53$] & [$-0.52,~~+0.61$]& [$-0.52,~~+0.62$] & [$-0.67,~~+0.62$] \\ 
896: $f_5^Z$      & [$-0.77,~~+0.86$] & [$-0.44,~~+0.69$]& [$-0.47,~~+1.39$] & [$-0.95,~~+0.33$] \\ 
897: \hline
898: \end{tabular}
899: \caption[]{The 95\% C.L. intervals ($\Delta\LL=1.92$) measured by
900:   ALEPH, DELPHI, L3 and OPAL.  In each case the two parameters listed are
901:   varied while the remaining ones are fixed to their Standard Model
902:   values.  Both statistical and systematic uncertainties are included.
903:   }
904: \label{tab:gc_fTGC-2-ADLO}
905: \end{center}
906: \end{table}
907: 
908: \begin{table}[htbp]
909: \begin{center}
910: \renewcommand{\arraystretch}{1.3}
911: \begin{tabular}{|l||c|rr|} 
912: \hline
913: Parameter     & 95\% C.L. & \multicolumn{2}{|c|}{Correlations} \\
914: \hline
915: \hline
916: $f_4^\gamma$  &[$-0.17,~~+0.19$] & $ 1.00$ & $ 0.07$\\ 
917: $f_4^Z$       &[$-0.30,~~+0.29$] & $ 0.07$ & $ 1.00$\\ 
918: \hline
919: $f_5^\gamma$  &[$-0.34,~~+0.38$] & $ 1.00$ & $-0.17$\\ 
920: $f_5^Z$       &[$-0.38,~~+0.36$] & $-0.17$ & $ 1.00$\\ 
921: \hline
922: \end{tabular}
923: \caption[]{ The 95\% C.L. intervals ($\Delta\LL=1.92$) obtained
924:   combining the results from all four experiments.  In each case the
925:   two parameters listed are varied while the remaining ones are fixed
926:   to their Standard Model values.  Both statistical and systematic
927:   uncertainties are included. Since the shape of the log-likelihood is
928:   not parabolic, there is some ambiguity in the definition of the
929:   correlation coefficients and the values quoted here are approximate.
930:   }
931:  \label{tab:gc_fTGC-2-LEP}
932: \end{center}
933: \end{table}
934: 
935: \clearpage
936: 
937: \begin{figure}[p]
938: \begin{center}
939: \includegraphics[width=\linewidth]{gc_aa03_fiv.eps}
940: \caption[]{
941:   The $\LL$ curves of the four experiments, and the LEP combined curve
942:   for the four neutral TGCs $f_i^V,~V=\gamma,Z,~i=4,5$.  In each case,
943:   the minimal value is subtracted.  }
944: \label{fig:gc_fTGC-1}
945: \end{center}
946: \end{figure}
947: 
948: \begin{figure}[p]
949: \begin{center}
950: \includegraphics[width=0.55\linewidth]{gc_aa03_f4gf4z.eps}\\
951: \includegraphics[width=0.55\linewidth]{gc_aa03_f5gf5z.eps}
952: \caption[]{
953:   Contour curves of 68\% C.L. and 95\% C.L. in the plane
954:   $(f_4^\gamma,f_4^Z)$ and $(f_5^\gamma,f_5^Z)$ showing the LEP
955:   combined result.  }
956: \label{fig:gc_fTGC-2}
957: \end{center}
958: \end{figure}
959: 
960: \clearpage
961: 
962: \subsection{Quartic Gauge Boson Couplings}
963: The individual numerical results from the experiments participating in the
964: combination, and the combined result are shown in Table~\ref{tab:gc_QGCs}. 
965: The corresponding $\LL$ curves are shown in
966: Figure~\ref{fig:gc_QGC-1}. The errors include both statistical and
967: systematic uncertainties. 
968: 
969: \begin{table}[ht]
970: \begin{center}
971: \begin{tabular}{|l||c|c|c|c|} \hline
972: Parameter & ALEPH  & L3  & OPAL   & Combined  \\
973: \hline \hline
974: \acl  &  [$-0.041,~~+0.044$]  &  [$-0.037,~~+0.054$] & 
975:          [$-0.045,~~+0.050$]  &  [$-0.029,~~+0.039$]  \\  
976: \azl  &  [$-0.012,~~+0.019$]  &  [$-0.014,~~+0.027$] & 
977:          [$-0.012,~~+0.031$]  &  [$-0.008,~~+0.021$] \\ 
978: \hline
979: \end{tabular}
980: \end{center}
981: \caption{ The limits for the QGCs \acl\ and \azl\ associated
982:     with the \ZZgg\ vertex at 95\% confidence level for ALEPH, L3 and OPAL, 
983:     and the LEP result obtained by combining them. 
984:     Both statistical and systematic errors are included.
985:   }
986: \label{tab:gc_QGCs}
987: \end{table}
988: 
989: \begin{figure}[htbp]
990: \begin{center}
991: \includegraphics[width=\linewidth]{gc_qgc1d_aachen03.eps}
992: \caption[]{
993:   The $\LL$ curves of L3 and OPAL (thin lines) and the 
994:   combined curve (thick line) for the QGCs \acl\ and \azl, associated with
995:   the \ZZgg\  vertex.  In each case, the minimal value is subtracted.  } 
996: \label{fig:gc_QGC-1}
997: \end{center}
998: \end{figure}
999: 
1000: 
1001: \section*{Conclusions}
1002: Combinations of charged and neutral triple gauge boson couplings, as well as
1003: quartic gauge boson couplings associated with the \ZZgg\ vertex were made,
1004: based on results from the four LEP experiments ALEPH, DELPHI, L3 and OPAL.
1005: No significant deviation
1006: from the Standard Model prediction is seen for any of the electroweak
1007: gauge boson couplings studied. 
1008: With the LEP-combined charged TGC results,
1009: the existence of triple gauge boson couplings among the electroweak
1010: gauge bosons is experimentally verified.
1011: As an example, these data allow
1012: the Kaluza-Klein theory~\cite{gc_bib:klein}, in which $\kg = -2$, to be
1013: excluded completely~\cite{gc_bib:maiani}.
1014: 
1015: 
1016: %
1017: %
1018: %
1019: %
1020: 
1021: %
1022: %
1023: %
1024: %
1025: