9 July 1998

Ž .Physics Letters B 431 1998 179–187

Determination of the Michel parameters r, j , and d

in t-lepton decays with t™rn tags

ARGUS Collaboration

H. Albrecht a, T. Hamacher a, R.P. Hofmann a, T. Kirchhoff a, R. Mankel a,1,A. Nau a, S. Nowak a,2, D. Reßing a, H. Schroder a, H.D. Schulz a, M. Walter a,2,¨

R. Wurth a, C. Hast b, H. Kapitza b, H. Kolanoski b,1, A. Kosche b, A. Lange b,A. Lindner b, M. Schieber b, T. Siegmund b, H. Thurn b, D. Topfer b,¨

D. Wegener b, C. Frankl c, J. Graf c, M. Schmidtler c,3, M. Schramm c,K.R. Schubert c, R. Schwierz c, B. Spaan c, R. Waldi c, K. Reim d, H. Wegener d,

R. Eckmann e, H. Kuipers e, O. Mai e, R. Mundt e, T. Oest e, R. Reiner e,A. Rohde e, W. Schmidt-Parzefall e, J. Stiewe f, S. Werner f, K. Ehret g,

W. Hofmann g, A. Hupper g, K.T. Knopfle g, J. Spengler g, P. Krieger h,4,¨ ¨D.B. MacFarlane h,5, J.D. Prentice h,4, P.R.B. Saull h,5, K. Tzamariudaki h,5,

R.G. Van de Water h,4, T.-S. Yoon h,4, M. Schneider i, S. Weseler i, M. Bracko j,ˇj j j j,6 j ˇ jG. Kernel , P. Krizan , E. Kriznic , G. Medin , T. Podobnik , T. Zivko ,ˇ ˇ ˇ

V. Balagura k, S. Barsuk k, I. Belyaev k, R. Chistov k, M. Danilov k, V. Eiges k,L. Gershtein k, Yu. Gershtein k, A. Golutvin k, O. Igonkina k, I. Korolko k,G. Kostina k, D. Litvintsev k, P. Pakhlov k, S. Semenov k, A. Snizhko k,

I. Tichomirov k, Yu. Zaitsev k

a DESY, Hamburg, Germanyb Institut fur Physik 7 UniÕersitat Dortmund, Germany¨ ¨

c Institut fur Kern- und Teilchenphysik 8 Technische UniÕersitat Dresden, Germany¨ ¨d Physikalisches Institut 9, UniÕersitat Erlangen-Nurnberg, Germany¨ ¨e II. Institut fur Experimentalphysik, UniÕersitat Hamburg, Germany¨ ¨

f Institut fur Hochenergiephysik 10, UniÕersitat Heidelberg, Germany¨ ¨g Max-Planck-Institut fur Kernphysik, Heidelberg, Germany¨

h Institute of Particle Physics 11, Canadai Institut fur Experimentelle Kernphysik 12, UniÕersitat Karlsruhe, Germany¨ ¨

j Institut J. Stefan and Oddelek za fiziko 13, UniÕerza Õ Ljubljani, Ljubljana, SloÕeniak Institute of Theoretical and Experimental Physics 14, Moscow, Russia

Received 8 December 1997Editor: K. Winter

0370-2693r98r$ – see frontmatter q 1998 Elsevier Science B.V. All rights reserved.Ž .PII: S0370-2693 98 00565-6

( )H. Albrecht et al.rPhysics Letters B 431 1998 179–187180

Abstract

Using the ARGUS detector at the eqey storage ring DORIS II, we have measured the Michel parameters r, j , and jd" "for t ™ l nn decays in t-pair events produced at center of mass energies in the region of the F resonances. Using

t .™r .n as spin analyzing tags, we find r s0.68"0.04"0.08, j s1.12"0.20"0.09, jd s0.57"0.14"0.07,e e e

r s0.69"0.06"0.08, j s1.25"0.27"0.14 and jd s0.72"0.18"0.10. In addition, we report the combinedm m m

ARGUS results on r, j , and jd using this work und previous measurements. q 1998 Elsevier Science B.V. All rightsreserved.

1. Introduction

w xA general ansatz to describe purely leptonic t decays is the well-known formalism by Michel 1 . Thisformalism includes extensions of the Standard Model such as scalar and tensor coupling to left- or righthandedleptons. The derived differential width in the t rest frame neglecting radiative corrections and termsproportional to m2rm2 isl t

2 " " 2 5d G t ™ l nn G m 2 m 1yxŽ . F t l2s x P 3 1yx q r 4 xy3 q6hŽ . Ž .4dV dx 3 m x192p t

2"P cosq j 1yx q jd 4 xy3 , 1Ž . Ž . Ž .t ž /3

where xs2 E rm is the scaled energy of the charged lepton, P the t polarization and q the angle betweenl t t

Ž .the polarization vector and the charged lepton momentum. Eq. 1 is valid for vanishing neutrino masses. TheStandard Model with a pure VyA interaction predicts rs3r4, hs0, js1, and jds3r4.

Up to now, measurements of the Michel parameters j and jd have been made by ARGUS, ALEPH, L3,w x w xSLD and CLEO 2–4 . In this paper, we present measurements of r, j , and jd with a new tagging method 5

separately for electrons and muons. The results are obtained using tau pairs produced in eqey™tqty™Ž " .Ž . . " "r n l n n . The decay t ™r n is used as a tag, i.e. for the event selection and as t spin analyzer.t l t t

w xThey are statistically independent of previous ARGUS measurements using lepton tags 2 and a -meson tags in1w xits decay into three charged pions 3 .

1 Now at Institut fur Physik, Humboldt-Universitat zu Berlin, Germany.¨ ¨2 DESY, IfH Zeuthen, Germany.3 Now at Caltech, Pasadena, USA.4 University of Toronto, Toronto, Ontario, Canada.5 McGill University, Montreal, Quebec, Canada.6 On leave from University of Montenegro, Yugoslavia7 Supported by the German Bundesministerium fur Forschung und Technologie, under contract number 054DO51P.¨8 Supported by the German Bundesministerium fur Forschung und Technologie, under contract number 056DD11P.¨9 Supported by the German Bundesministerium fur Forschung und Technologie, under contract number 054ER12P.¨

10 Supported by the German Bundesministerium fur Forschung und Technologie, under contract number 055HD21P.¨11 Supported by the Natural Sciences and Engineering Research Council, Canada.12 Supported by the German Bundesministerium fur Forschung und Technologie, under contract number 055KA11P.¨13 Supported by the Ministry of Science and Technology of the Republic of Slovenia and the Internationales Buro KfA, Julich.¨ ¨14 Partially supported by Grant MSB300 from the International Science Foundation.

( )H. Albrecht et al.rPhysics Letters B 431 1998 179–187 181

2. Method of the measurement

Ž .As can be seen from Eq. 1 , the measurement of j and jd requires the knowledge of the t polarization. AtDORIS energies, the average polarization of t leptons is zero, hence no information on j and jd can beextracted from the analysis of single t-decays. Only r and h are accessible from the analysis of the momentum

'spectrum. However, t-leptons are produced in pairs resulting in a spin-spin-correlation. At s s 10 GeV, t

pairs are produced via a virtual photon and both t spins are predominantly parallel due to the vector coupling,Ž . 2with only a small probability of 1yb r2fm rsf3% for being antiparallel. In the analysis described heret t

w x " "6 , we use the decay t ™r n in order to obtain information on the t-spin orientation. The basic method ist" " w xsimilar to the one used in the ARGUS analysis with t ™a n as spin analyzer 3 , and is described in more1 t

detail there.The performance of various analysis methods has been studied using a Monte Carlo simulation. Tau pair

w xevents including radiative photons were produced by the KoralBrTauola generator 7 . These events passed thew x w xARGUS detector simulation Simarg 8 and the reconstuction program Arg13 9 , and were finally submitted to

the same selection procedure as the data, which is described in Section 3 below."The parallel orientation of the two t spins is used to relate the spin of the tau decaying into l nn to the spin

of the tau decaying into r "n . The spin in the latter decay is related to the r " direction because of parityt" " w xviolation with a neutrino helicity h sy1 as supported by measurements of the sign in t ™a n decays 3n 1 tt

q y q yŽ .Ž . w xand the modulus in t t ™ r n r n events 10 . The r-meson can be produced in two helicity states,t t

Ž ) . Ž ) . 2 2hs0 and hsq1. The state with hs0 dominates by a factor of 1qb r 1yb sm rm f5.3, wheret r

b ) is the velocity of the t in the r rest frame. Therefore, its t-spin r-momentum correlation also dominates,resulting in correlations between the momenta p of the lepton and p of the r-meson of the same tau pairl r

event. In a Monte Carlo sample with the Standard Model value js1 on the lepton side and h sy1 on the rnt

Ž . Ž Ž . Ž .. Ž .side, we obtain a correlation coefficent rsCov p , p r s p Ps p s y6.6"0.4 % whereas parityl r l r

Ž . Ž . q yconservation h Pjs0 leads to rs y2.2"0.4 % which is non-zero because of radiation in the initial e ent

state. A measurement of r is, therefore, a measurement of h Pj .nt

Ž .Additional information on j and jd is obtained by i using the momentum-momentum correlations of theŽ .vectors instead of just their absolute values, and ii including information on the r-meson helicity, since the

suppressed hsq1 state results in the opposite correlation compared to the hs0 state. A discriminationbetween both states is possible via the decay angle for r "™p "p 0, since the distributions for both helicitiesare different.

Maximal information on r, j , and jd is obtained by a single entry maximum likelihood fit using the" . 0 q y 'Ž .Ž .selected l nn p p n events with all their 10 measured quantities, the e e cms energy s , and the

Ž .. Ž ". Ž 0.momentum vectors of the three observed particles, p l , p p , and p p . The matrix element for theq y q y " . 0Ž .Ž .differential cross section e e ™t t ™ l nn p p n can be written as

2 X X Xa b< < w xMM sH A L qrL qhL qh H C j L qjdL , 2Ž .1 2 3 n a 1b 2 bt

where the first part is the spin-averaged contribution and the second part describes the tqty-spin correlations. HX X X" " " " 0 " "and H describe the decays t ™r n and r ™p p . L , L , L , L , and L describe t ™ l nn decays1 2 3 1 2

in the most general way with zero neutrino masses. A is the spin-averaged matrix element for t-pair productionŽ . < < 2and C the spin-spin correlation matrix a ,bs1 . . . 4 . MM includes radiative corrections which are treated in

a factorizable way.Ž .The likelihood function to be maximized contains three variables us r,j ,jd and NP10 measured

Ž Ž .. Ž ". Ž 0..quantities m s s , p l , p p , p p , where is1 . . . N, and N is the number of selected events,i i i i i

N N

< <L u s L u m s f m u . 3Ž . Ž .Ž . Ž .Ł Łi i iis1 is1

( )H. Albrecht et al.rPhysics Letters B 431 1998 179–187182

Since data were taken at different cms energies, the probability density f is taken as a function of scaled' Ž < .momenta p r s . It is the projection of the full density f m,u u where u denotes the unobserved kinematicali

quantities which include neutrino momenta as well as photon momenta of initial and final state radiation.Furthermore, the observables m are obtained as a convolution of the true values t of the same kinematicali i

quantities and the detector resolution function RR. The final density is given by

< < < < 2 <f m u sh m PNN u MM t ,u u PPP t ,u PRR m t dt du , 4Ž . Ž . Ž . Ž . Ž . Ž . Ž .HŽ . Ž .where PP is the phase space distribution, h m is the detector acceptance, and NN u is the normalization factor

given by

< < < 2 <NN u P h m P MM t ,u u PPP t ,u PRR m t dt du dms1. 5Ž . Ž . Ž . Ž . Ž . Ž .HRadiative corrections including initial state radiation and internal bremsstrahlung in t™enn are included inŽ < . Ž < .MM t,u u . External bremsstrahlung of the electron is included in the resolution function RR m t . For the

Ž < .application as a likelihood function, f m u is rewritten in a more appropriate form. The factorizing radiative< Ž < . < 2 Ž . 3 Ž .corrections allow to write MM t,u u PPP t,u sÝ u F t,u with u '1. The likelihood of a single event ijs0 j j 0

is then3

<u F t ,u RR m t dt duŽ . Ž .Ý Hj j ijs0

<L u m s , 6Ž .Ž .i i 3

<u h m F t ,u RR m t dt du dmŽ . Ž . Ž .Ý Hj jjs0

where the acceptance factor has been omitted in the numerator since it depends only on the observed quantitiesm. The integration over the neutrino momenta, which are part of the unobserved quantities u, is performed to alarge extent analytically, leaving only an uncertainty in the relative angular orientation of the two tau leptons.The integration over t and the remaining seven unobserved variables u is done by a Monte Carlo method in

" . 0Ž .Ž .Fig. 1. Number of matching e nn p p n configurations per event. The points with error bars represent the data. The hatched" . 0Ž .Ž .histogram shows the expectation from the KoralBrTauola Monte Carlo for the decay e nn p p n and other contributing t channels.

The solid line marks n s25, the minimum number for selected events. Events with a hard initial state photon lead mostly to n -25.hit hit

( )H. Albrecht et al.rPhysics Letters B 431 1998 179–187 183

which for each event 450 kinematical configurations are tested for compatibility with the observed set ofobservables m . The seven u variables in the Monte Carlo integration are the momentum vector of the initiali

state photon, the t orientation angle, and the momentum vector of the photon in t™enng . A configuration is2q y " 0 .Ž . Ž .Ž .‘‘successful’’ if the set of m ,t,u is compatible with the kinematics of e e ™g p p n l nng events,i 1 2

i.e. if the undetermined neutrino momenta can have physical values.These 450 tries lead to n successful matches for each event, and the distribution of n is shown in Fig. 1.hit hit

² :The high value of n shows the effectiveness of the numerical integration chosen, and n is large enoughhit hitŽ < .for most events to compute L u m with sufficient precision. Events with n -25 are rejected in the datai i hit

Ž .selection. The good agreement between N n for selected data and for accepted Monte Carlo events supportshit

the validity of the assumptions used.Ž .The integral in the denominator of Eq. 6 , for each set of parameters u , is also obtained with the help of a

Monte Carlo integration which determines the four separate integrals used in the sum. The validity of themethod has been extensively tested by using Monte Carlo events as pseudo data for determining the parameterr, j , and jd . Even using pseudo data samples 10 times larger than the real data samples, we did not detect anysignificant biases in the determined parameter values.

3. Data selection

The measurements presented here were performed with the ARGUS detector at the eqey-storage ringw xDORIS II. The detector and its trigger requirements are described elsewhere 9 . The data sample used was

collected between 1983 and 1989 in the center of mass energy region of the F resonances. The integratedluminosity used in this analysis is 289 pby1 corresponding to 265 000 t pairs produced.

Event selection starts with requiring two oppositely charged tracks forming a vertex in the interaction regionand an opening angle larger than 808 at the vertex point. Each track must have a transverse momentum above150 MeVrc and point into the barrel region. We demand one particle to be identified as a lepton. The other

) 'Fig. 2. Spectra of scaled momentum p sp P10 GeVr s for leptons. Data are shown as dots with error bars and accepted Monte Carlol l

events, generated with rs 3r4, as grey shaded area.

( )H. Albrecht et al.rPhysics Letters B 431 1998 179–187184

" 0 .Ž .Ž ." 0Fig. 3. Invariant mass distribution m for the p p n m nn signature. The Monte Carlo spectrum includes background from otherp p

tau decays.

track has to be accompagnied by one or two neutral clusters in its hemisphere with a minimum angle of 108 tothe track. A neutral cluster is defined as energy deposition above 100 MeV in the calorimeter. No additionalcluster is allowed, neither in the lepton hemisphere nor in the 108 cone around the p " candidate. In the singlecluster case, accounting for photons from the p 0 decay merging into one cluster, we demand a minimumenergy deposition of 1 GeV. In the two cluster case, both photons are combined to form a p 0. The two photonsystem must have an invariant mass within 100 MeV of the nominal p 0 mass and x 2 F 9 for the p 0 masshypothesis. The lepton identification was done by a likelihood method using specific energy loss, time of flight,amount and lateral spread of energy deposition in the calorimeter and hits in the muon chambers. More details

w xon this likelihood method are found in Refs. 11,12 . To ensure a good lepton identification, a momentum above0.8 GeVrc for electrons and above 1.1 GeVrc for muons as well as an electron or muon likelihood higher than0.8 is required. The charged pion candidate is required to have a pion likelihood above 0.6 using specific energy

w xloss in the drift chamber and time-of-flight 9 .To reject background, we apply the following three additional event cuts. The invariant mass m " 0 has top p

be smaller than the t lepton mass. This cut suppresses qq-events. The angle between the lepton momentum pl

and the r-meson momentum p is not larger than 1768, which rejects radiative Bhabha and m pair events. Ther2 4 'Ž .relation f P m z rsy f -p r s with f s5 and f s0.45 for the single cluster case, and f s3 andc miss o t c o c

f s0.4 for the two cluster case has to be fulfilled. In here, the missing mass m uses only charged tracks ando miss

the vector sum of transverse momenta p uses charged and neutral particles. This cut removes gg and radiativet

Bhabha events.

Table 1Ž .Monte Carlo determined background levels as used in the likelihood function Eq. 4 . The asymmetric errors are due to the systematic

deviation of ARGUS branching fractions from the world average used to calculate these factors" 0 . " 0 .Ž .Ž . Ž .Ž .p p n e nn p p n m nn

Ž . Ž .l 0.8"0.2 % 1.6"0.2 %f2.8 2.9Ž . Ž .l 10.1" % 10.5" %a 0.6 0.61

Ž . Ž .)l 3.7"1.0 % 3.4"1.0 %K

( )H. Albrecht et al.rPhysics Letters B 431 1998 179–187 185

Table 2Contributions to the systematic error

" . 0 " . 0Ž .Ž . Ž .Ž .systematic error source e nn p p n m nn p p n

Ž . Ž . Ž . Ž . Ž . Ž .D r D j D jd D r D j D jd

Monte Carlo statistics "0.01 "0.05 "0.03 "0.02 "0.07 "0.04Trigger efficiency "0.01 "0.01 "0.01 f0 "0.07 "0.01

q0 .020 q0.041 q0.024 q0.026 q0.044 q0.057Branching fractions y0 .003 y0.007 y0.004 y0.004 y0.007 y0.009

energy dependence "0.03 "0.02 "0.02 "0.03 "0.03 "0.02q0 .05particle identification "0.03 "0.03 "0.04 "0.02 "0.02y0 .06

lepton misidententification 0.005 "0.01 "0.004 0.005 "0.02 "0.005Ž .background non t physics "0.01 "0.03 "0.03 "0.01 "0.07 "0.06

total systematic error "0.08 "0.09 "0.07 "0.08 "0.14 "0.10

" . 0Ž .Ž .After these cuts, a data sample of 3176 candidates for e nn p p n and 2099 candidates for" . 0Ž .Ž .m nn p p n remains. As seen in Fig. 2, a good agreement for lepton momenta is found comparing data

and Monte Carlo events. However, the data selection criteria do not allow to remove all background sources." 0 0 . " 0 .Ž .Ž . Ž .Ž .The selected sample still contains 10% p p p n l nn decays, 3.5% K p n l nn decays, f 1%

misidentified leptons and around 2% background from non-t channels, as determined by Monte Carlo. Fig. 3shows the p "p 0 invariant mass distribution with good agreement between data and Monte Carlo events. Tau

w xdecay fractions for the background estimation are taken from Ref. 13 .

4. Data analysis

A first information on j is obtained from the p -p correlation coefficient. The selected data sample hasl r

Ž .rsy 6.1"1.2 % leading to js0.89"0.27 without background corrections. The finally used likelihood

Fig. 4. The contribution of single events to y2ln L) for events with n G 25. The points with error bars represent the data. The openi hit" . 0Ž .Ž .triangles show the expectation from the KoralBrTauola Monte Carlo for the decay e nn p p n and other contributing t channels.

The shaded area are Monte Carlo events without consideration of any background from other t decay channels or lepton misidentification.

( )H. Albrecht et al.rPhysics Letters B 431 1998 179–187186

Table 3Fitted Michel parameters

. 0 " . 0 "Ž .Ž . Ž .Ž .p p n e nn p p n m nn combined fit

r 0.68"0.04"0.07 0.69"0.06"0.06 0.68"0.04"0.07j 1.11"0.20"0.08 1.26"0.27"0.14 1.18"0.16"0.10jd 0.56"0.14"0.06 0.73"0.18"0.10 0.63"0.11"0.08

Ž .function in Eq. 6 can conveniently include background from other semihadronic t decays in the data sampleand from lepton misidentification. The modified likelihood function is

L) s 1yl yl ) 1yl PL ql PL f ak e ql PL " 0 0 ql ) PL " 0 , 7Ž .Ž .Ž . Ž .i a K f i f i a i ,Žp p p . K i ,ŽK p .1 1

Ž " 0 0. Ž " 0.)where l is the fraction of the p p p -tags, l is the fraction of the K p -tags and l considersa K f1

pions misidentified as leptons. Table 1 gives the used values for l , l , and l ) as determined by Montef a K1" 0 0 . " 0 .Ž .Ž . Ž .Ž .Carlo simulation. The additional likelihood functions L for ks p p p n l nn and K p n l nni,k

" 0 . f ak eŽ .Ž .are determined by the same technique as L for p p n l nn . L is the contribution to the likelihoodi i

function for misidentified leptons. Non-t background has not been considered in the likelihood function,because its contribution is small and is included in the systematic error studies, see Table 2.

The maximum likelihood fits lead to results and statistical errors for the Michel parameters as shown in Table3 below. Figs. 1 and 4 illustrate the quality of the fit and the validity of the assumptions made for thedetermination of the kinematical quantities used to calculate the matrix elements.

Fig. 1 shows the distribution of successful tries to find a matching initial state to the observed final statecomparing data and Monte Carlo events. Fig. 4 shows the agreement in y2ln L) for the fit to the data comparedi

Ž .to the fit to Monte Carlo events generated with Standard Model parameters described with Eq. 7 . Deviationsare found at large values of y2P ln L) if we neglect other t sources and misidentified leptons.i

The systematic errors are summarized in Table 2 including limited Monte Carlo statistics, trigger efficiencies,branching fractions, particle identification efficiencies, lepton misidentification, the influence of backgroundfrom non-t physics and the center of mass energy dependence of the matrix element description. The total erroris obtained as a quadratic sum of all errors in Table 2.

5. Results and conclusions

The Michel parameters as found by the maximum likelihood analysis described above, and includingsystematic errors, are given in Table 3.

The results are in agreement with previous measurements and the Standard Model predictions. In addition,the errors are comparable to the errors of previous measurements. The results with r constrained to the world

w xaverage of rs0.742"0.027 14 are very close to the unconstrained ones; they are given in Table 4 forcompleteness.

All quoted values for j and jd assume a t neutrino helicity h sy1. Using the experimental averagent

w xh sy1.011"0.027 14 would not change the results.nt

Assuming lepton universality and combining the results of this analysis in Table 3 with previous ARGUSw xmeasurements 15 of r, j , and jd , we obtain

rs0.731"0.031

js1.03"0.11

jds0.63"0.09

( )H. Albrecht et al.rPhysics Letters B 431 1998 179–187 187

Table 4Fitted results with constrained r

. 0 " . 0 "Ž .Ž . Ž .Ž .p p n e nn p p n m nn combined fit

r 0.72"0.02 0.73"0.02 0.72"0.02j 1.12"0.20"0.09 1.26"0.27"0.14 1.18"0.16"0.10jd 0.59"0.14"0.07 0.74"0.18"0.10 0.64"0.11"0.08

where statistical and systematical errors are combined in quadrature. Correlations in the systematic errors ofdifferent measurements are taken into account. The covariance matrix is given by

r j jd

y4r 9.54P10y5 y2j y2.30P10 1.25P10y5 y3 y3jd y7.88P10 1.03P10 7.84P10

Acknowledgements

It is a pleasure to thank U. Djuanda, E. Konrad, E. Michel, and W. Reinsch for their competent technicalhelp in running the experiment and processing the data. We thank Dr. H. Nesemann, B. Sarau, and the DORISgroup for the operation of the storage ring. The visiting groups wish to thank the DESY directorate for thesupport and kind hospitality extended to them.

References

w x Ž .1 L. Michel, Proc. Phys. Soc. London A 63 1950 514.w x Ž .2 ARGUS Collaboration, H. Albrecht et al., Phys. Lett. B 316 1993 608.w x Ž .3 ARGUS Collaboration, H. Albrecht et al., Phys. Lett. B 349 1995 576, M. Schmidtler, Bestimmung der Michelparameter j und d in

leptonischen t-Zerfallen, Dr. rer. nat. Thesis, Universitat Karlsruhe, IEKP-KAr94-16, 1994.¨ ¨w x Ž . Ž .4 ALEPH Collaboration, Phys. Lett. B 346 1995 379; L3 Collaboration, Phys. Lett. B 377 1996 313; SLD Collaboration, Phys. Rev.

Ž .Lett. 78 1997 4691; CLEO Collaboration, CLNS-97-1480, 1997.w x5 ARGUS Collaboration, contributed paper pa 07-098 to the 28th International Conference on High Energy Physics, Warsaw, Poland,

July 1996, TUD-IKTPr96-01.w x6 M. Schramm, Bestimmung der Michelparameter j und d in leptonischen t-Zerfallen mit r-Mesonen als Spinanalysator, Dr. rer. nat.¨

Thesis, Technische Universitat Dresden, TUD-IKTPr97-01, 1997.¨w x Ž .7 S. Jadach, J.H. Kuhn, Z. Wa̧s, Comp. Phys. Comm. 64 1991 275.¨w x8 H. Gennow, SIMARG - A Program to simulate the ARGUS Detector, DESY F15-85-02, 1985.w x Ž .9 ARGUS Collaboration, H. Albrecht et al., Nucl. Instr. Meth. A 275 1989 1.

w x Ž .10 ARGUS Collaboration, H. Albrecht et al., Phys. Lett. B 337 1994 383; ARGUS Collaboration, H. Albrecht et al., Z. Phys. C 58Ž .1993 61.

w x Ž .11 S. Weseler, Untersuchungen der semileptonischen Zerfalle von B 5270 -Mesonen mit dem ARGUS-Detektor, Dr. rer. nat. Thesis,¨Universitat Heidelberg, IHEP-HDr86-02, 1986.¨

w x12 B. Fominykh, V. Matveev, ARGUS MUON Program, ARGUS software note 14, April 1987.w x Ž .13 Particle Data Group, L. Montanet et al., Phys. Rev. D 50 1994 1173 and 1995 off-year update for the 1996 edition.w x Ž .14 Particle Data Group, R.M. Barnett et al., Phys. Rev. D 54 1996 1.w x Ž .15 ARGUS Collaboration, H. Albrecht et al., Phys. Rep. 276 1996 223.