= SPEAR Mark I tau discovery experiment
{c}
{title2=1975}
Counterrotating <Electron> and <Positron> beams at SPEAR collided inside the Mark I magnetic <particle detector>. Lead–scintillator <scintillation counters> identified <Electrons> through large shower pulses; small pulses plus hits beyond the iron return identified <muons>.<Evidence for Anomalous Lepton Production in e+ e- Annihilation by Martin Lewis Perl and others>[SLAC-PUB-1626, pp. 2–3]{ref}
The selection required opposite-charge <Electron>–<muon> tracks, no detected <photons>, <momenta> above $0.65\,\mathrm{GeV}/c$, and transverse noncoplanarity above $20^\circ$. Multitrack <hadron> samples calibrated momentum-dependent misidentification; collinear <lepton> pairs checked identification losses. Photon and charge control samples tested backgrounds. Acceptance corrections depended on the proposed signal.<Evidence for Anomalous Lepton Production in e+ e- Annihilation by Martin Lewis Perl and others>[SLAC-PUB-1626, pp. 3–5, Table I]{ref}
The 1975 claim was anomalous events without a conventional explanation; new-particle pair production was a possible interpretation.<Evidence for Anomalous Lepton Production in e+ e- Annihilation by Martin Lewis Perl and others>[Abstract; SLAC-PUB-1626, pp. 5–6]{ref} <Martin Lewis Perl>'s lecture describes subsequent evidence for a <charged lepton>.<Reflections on the Discovery of the Tau Lepton by Martin Lewis Perl>[pp. 184–187]{ref}
\Image[/tau-electron-muon-event-geometry.png]
{title=An electron–muon signature and its invisible recoil}
{description=Functional identification paths beside an original, conserved event in a massless-daughter teaching model. Solid arrows show visible transverse momenta; the dashed arrow is the sum of four neutrino momenta in this chosen interpretation. It is neither a measured track nor a historical event.}
The <particle identification> problem matters as much as the apparent <missing four-momentum in electron-positron collisions>. A penetrating <hadron> can imitate a <muon>; a <particle shower> can imitate an <Electron>; an undetected <photon> or a detector gap can imitate invisible recoil. An absence of additional recorded signals is conditional on <detection efficiency>. In a proposed signal model, the selected count has the schematic form
$$
N_{\rm selected}=\mathcal L\,\sigma\,\varepsilon+N_{\rm background},
$$
where $\mathcal L$ is integrated <collider luminosity>, $\sigma$ a <cross section>, and $\varepsilon$ the combined acceptance and <detection efficiency>. A single count cannot determine all three factors. A control sample must probe the relevant <momentum> and angular range; its abundant events do not automatically calibrate a differently distributed signal.
Back to article page
