Main-sequence photometry, color-magnitude diagrams, and ages for the globular clusters M3, M13, M15, and M92

New photoelectric photometry is reported for four globular clusters. Distance moduli were found by using nearby field subdwarfs with trigonometric parallaxes for calibration, with account taken of the ultraviolet excess. If one excludes M13, whose RR Lyrae stars may be related to the asymptotic branch rather than the horizontal branch, the RR Lyrae stars in M3, MiS, and M92 have (4iv) +0.6 + 0.2 by this method. The mean is close to values obtained by Woolley et at. and van Herk from statistical parallaxes, and by Christy from pulsation theory. Christy's calculations of absolute magnitudes are consistent with the Oosterhoff-Sawyer effect, and are adopted as a more precise method to obtain luminosities at the termination of the main sequence. Final adopted moduli are (ni - )o,Mi = 14.83, (m - AI)Av,M13 = 14.42, (m - I) = 14.93, and (m - I)Av,M6i = 14.63, where the subscripts zero and A V mean true and apparent visual moduli, respectively. The adopted turnoff luminosities and the helium and metal abundances used for age dating are listed in a table. Ages calculated from Iben and Rood's models average 11.5 X l0 years and have a very small spread. Ages from models by Demarque et at. are about 20 percent smaller. These ages show no major conflict with limits for the Friedmann time of the expanding Universe when the present uncertainties in the Hubble constant and the deceleration parameter are considered. The wel known correlation of the Oosterhoff-Sawyer period-groups with metal ahundance is shown to be a natural consequence of the prediction of Simoda and Iben that for clusters of equal age, those with lower metal abundance have brighter luminosities at main-sequence turnoff. The observations and theory are combined to show that the most probable age spread among the four clusters is A%/ 0.014 or AT 1.5 X 108 years as required in the rapid-collapse picture of the Galaxy discussed by Eggen, Lynden-Bell, and Sandage. But the limiting error of the result is high, and permits limits of about AT l0 years for the age spread. Representative points of the C-M diagrams for the four clusters are tabulated in the Appendix.