It has long been assumed that hole self-trapping should take place in amorphous silicon dioxide (a-SiO2). However, no spectroscopic evidence for this was claimed before 1989, when the author used electron spin resonance (ESR) to identify self-trapped holes (STHs) in bulk samples of low-OH pure fused silica X irradiated ⩽100K, and Chernov et al. reported a low-temperature infrared absorption near 1600nm in irradiated pure-silica-core fibers, which they ascribed to STHs. Based on g values and 29Si and 17O hyperfine coupling constants measured by ESR, Griscom [D.L. Griscom, Phys. Rev. B 40 (1989) 4224; D.L. Griscom, J. Non-Cryst. Solids 149 (1992) 137] deduced the existence of two types of STHs: STH1 (a hole trapped on a single bridging oxygen) and STH2 (a hole delocalized over two equivalent bridging oxygens of the same SiO4 tetrahedron). The validity of Griscom’s models for STH1 and STH2 are supported by the ab initio calculations of Pacchioni and Basile [G. Pacchioni, A. Basile, Phys. Rev. B 60 (1999) 9990] and Gabriel [M.A. Gabriel, PhD dissertation, Department of Chemistry, University of Washington, Seattle, WA, in preparation]. In 1984, Nagasawa et al. reported that low-OH-pure-silica-core optical fibers γ irradiated at ∼300K exhibit metastable optical absorption bands at 660 and 760nm. Griscom [D.L. Griscom, J. Non-Cryst. Solids 349 (2004) 139] recorded these same bands in low-OH-pure-silica-core fibers γ irradiated at 77K, showing their isochronal annealing behaviors to correlate with his earlier ESR data for STHs in bulk silica and also with Harari et al.’s data [E. Harari, S. Wang, B.S.H. Royce, J. Appl. Phys. 46 (1975) 1310] for trapped positive charges in silica thin films following X irradiation at 77K. Sasajima and Tanimura [Y. Sasajima, K. Tanimura, Phys. Rev. B 68 (2003) 014204] established direct correlations of an induced band at 574nm with STH2 by performing both ESR and a variety of optical measurements on three types of high-purity bulk silicas following pulsed electron irradiations at 77K; however, these authors did not detect the 660 or 760nm bands. Yamaguchi et al. [M. Yamaguchi, K. Saito, A.J. Ikushima, Phys. Rev. B 68 (2003) 153204] demonstrated that the yield of ESR-detected STHs photoinduced in bulk pure-silica samples at 77K depends exponentially on fictive temperature (Tf). The present paper recounts the forgoing history in greater detail, while attempting to reconcile some seemingly disparate findings into a unified picture of STHs in silica. In the fall of 1998, a number of satellites in orbit were tumbling out of control due to failure of their HeNe ring-laser-gyro (RLG) attitude control systems. The author, acting in the capacity of pro-bono US government consultant, proposed the correct solution to this problem (replace Al-contaminated silica mirror coatings with high-purity ones). However, accelerated tests (short operation times at much-higher-than-normal laser powers) of the corrected devices failed to corroborate this fix. Thus, all further launches of satellites employing these RLGs remained grounded until the author was able to convince industry troubleshooters that (1) the accelerated test failures were due to the 660-nm STH band (which is induced even in the highest-purity silica coatings by 20-eV photons emitted by the laser plasma), (2) the strength of this band is initially proportional to ionizing dose rate (inevitably giving false positives in accelerated tests), and (3) this band eventually disappears after several months of irradiation, even at dose rates as low as 0.15Gy/s. All of these insights derived from curiosity-driven components of the author’s research.