... B-hadrons^1.1

In other words, the decay of naked beauty to hidden charm.

.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.

...Z-psi-upsilon^1.2

Reference [7] quotes a branching ratio of $3.3 \E{-4}$, which is based on an amplitude calculated in [9]. Using their updated amplitude from [10] gives $1.9 \E{-4}$.

.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.

... DELPHI^2.1

Detector with Lepton, Photon, and Hadron Identification, ©Gerald Myatt. The members of the DELPHI Collaboration are listed in appendix A.

.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.

... LEP^2.2

Large Electron-Positron collider

.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.

...^2.3

European Laboratory for Particle Physics. The acronym, `CERN', is from its original title, Conseil Européen pour la Recherche Nucléaire.

.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.

... peak^2.4

$1\ \ensuremathbox{\mathrm{nb}} = 10^{-33}\ \ensuremathbox{\mathrm{cm}}^2$

.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.

... SIROCCO^2.5

Silicon Strip Readout CAMAC Controller, anachronisticly named since they are Fastbus modules.

.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.

... VDCLAP^2.6

Vertex Detector Common Library for All Programs

.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.

...^2.7

The FEB is connected to the CEB. The CEB is connected to the SEB. The SEB is connected to the MEB. The MEB is connected to the GEB. The GEB is connected to the MBM. Hear the word of the Lord.

.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.

...TANAGRA^2.8

Track Analysis and Graphics

.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.

...^2.9

``TANAGRA is like the mafia: it protects you, it is expensive, and you did not ask for it.'' [55]

.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.

...DST^2.10

Data Summary Tapes

.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.

...^2.11

Note that unlike most other LEP and planned future detectors, the DELPHI simulation program is not based on GEANT [79], though GEANTH is used to simulate nuclear interactions.

.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.

...VD-cap^3.1

Until the end of 1993.

.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.

...^3.2

The elevated temperature was originally intended to also allow a higher gas radiator pressure (which would otherwise condense), providing improved pion/kaon differentiation at intermediate momenta. However this option has been ruled out for mechanical reasons.

.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.

...^3.3

Of these, 7 (6 supervisors and one supply monitoring) are not included in table 3.2.

.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.

...sec:SC-FCH^3.4

A general-purpose automatic trip-recovery system, drawing on experience from these modificaions, is now implemented in the standard Elementary Process.

.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.

...^3.5

The STIC can in fact be switched on after acceleration is complete, but before the collimators are closed. Doing this makes an additional measure of the background conditions available to the LEP and DELPHI operators.

.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.

...^3.6

Prior to the autumn of 1993 the high voltages were lowered before the beam was dumped using the Prepare_For_Dump command. We have since determined that the beam dump procedure does not produce any additional background radiation, so Prepare_For_Dump is now only used prior to a period of LEP machine development following a physics run.

.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.

...chapter.^3.7

The majority of the improvements were in operation at the start of datataking in 1992.

.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.

...^3.8

Buy your Slow Controls system from DELPHI! Washes whiter than white! (i.e. Whitewashes all known problems).

.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.

... DLT2000^4.1

digital linear tape

.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.

...^4.2

The Atlas Data Store is a facility to store large amounts of data, logically accessed as tape volumes of user-specified size, independently of their physical location in the Atlas Centre robots. Data can be accessed from on-site or from anywhere in the United Kingdom via SuperJANET.

.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.

...^4.3

By constraining the two tracks to go through the same point, the errors on the two particles' momenta become correlated. In principle this should be taken into account when combining them to form the errors on the parent's mass and momentum. A procedure for using the track-track correlations was developed. Since the change in the calculated errors turned out to be small ($\sim 3\%$ on the $\ensuremath{M_{\mu\mu}}$ error) and not noticeably better, for simplicity this procedure was disabled.

.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.

...^4.4

Routines developed for this analysis are now part of the DELPHI-standard DSTANA library [143]. They allow one to iterate over the vertex fit (VDF2ND), convert the track parameters to Cartesians (VDPE2E), calculate the parent's parameters (REC2ND), and from them its mass, momentum, and direction (VDMPER). Although only two daughters of equal mass are considered in this analysis, the code allows for an arbitrary number of tracks with individually assigned masses. At each stage the parameter errors are calculated, along with their correlations (though only optionally for the track-track correlations).

.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.

....^4.5

Actually $\mathbf{p}_{\ensuremathbox{\mathrm{S}}} \times \mathbf{p}_{\ensuremathbox{\mathrm{L}}}$ is used, where $\mathbf{p}_{\ensuremathbox{\mathrm{S}}}$ ( $\mathbf{p}_{\ensuremathbox{\mathrm{L}}}$) is the momentum of the track with the smaller (larger) geometric curvature ( $-qB/p_{xy}$) in the magnetic field. This definition works for both like- and unlike-sign candidates.

.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.

...^4.6

Simulated $\ensuremathbox{\mathrm{J/\psi}}$s that failed those cuts are nevertheless stored on the n-tuple, so must be rejected here ( $\mathtt{NUMB} > 0$).

.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.

...^4.7

To perform the fit within PAW on VMS, the Fortran fit function would have to be calculated using the COMIS [147] interpreter. While a wonderful tool for simple routines, it is considerably slower than a compiled program, making it impracticable as an interactive tool for unbinned fits with many data points.

.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.

...^4.8

A better description for the reconstructed mass pulls would be to include a second Gaussian of width 3.5 consisting of 2.6% of the data.

.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.

...^4.9

A Kolmogorov test was also used, but only as a crosscheck. Since the comparison is with the result of a fit rather then some theoretical function, Kolmogorov probabilities very close to 1 were obtained. However this might indicate a problem if a low probability were obtained.

.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.

...^4.10

$\Theta(x)$ is the Heaviside step function, which takes a value of zero for $x<0$ and unity for $x>0$. Its differential is the Dirac delta function: $\delta(x) \equiv \frac{\d {}}{\d {x}}\Theta(x)$.

.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.

...^4.11

Another possibility would be to perform a primary vertex fit for each event. However this gives little improvement in the overall lifetime fit (where the number of signal events is of greatest importance) and risks (slightly) reducing the sample size (due to events where no primary vertex fit can be performed) and introducing systematic effects (due to erroneously including secondary tracks in the primary vertex fit).

.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.

...^4.12

requiring triple integration using numeric methods to obtain the normalized function

.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.

...^4.13

Note that 28% ($\sim 59$ events) of the data that this average is based on is common with the 22% of our sample that was taken in 1992.

.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.

...DELPHI-psi^4.14

73% of its data is common with 22% of our sample.

.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.

...PDG96^4.15

10% of its data is common with 22% of our sample. The average does not include the recent L3 measurement [152] of $\ensuremathbox{\mathrm{Br}}(\ensuremathbox{\mathrm{Z^0}}\ensuremathbox{\righta... ...\ \ensuremathbox{\mathrm{(stat.)}} \pm 0.027\ \ensuremathbox{\mathrm{(syst.)}})$%, though perhaps J and $\ensuremathbox{\mathrm{J/\psi}}$ measurements should not be combined!

.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.

...fig:perigee).^B.1

It is common (eg. [66]) to define the sign of $\epsilon$ in terms of the track curvature (whether the origin is enclosed by the track), with an additional sign for a positively charged particle. Although equivalent to the one given above, such a definition engenders the incorrect implication that the track's charge plays a rôle, whereas in fact $\epsilon$ is equally applicable to neutral tracks.

.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.

... formula^D.1

[156, page III.36, equation 2.9] gives this incorrectly as $\chi^2 \equiv -\frac{1}{2} \ln{\cal L}_{\chi^2}(\mathbf{p})$, so maybe it isn't all that well known! (It is corrected in [157, equation 17.10].)

.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.