A detailed model atmosphere and chemical composition analysis has been made of the brightest known anomalous Cepheid, V19 in the globular cluster NGC 5466. Our study is based on 30 minute CCD echelle spectrograms (4300 less than or equal to lambda less than or equal to 6630 Angstrom) acquired in 1995 and 1996 with the HIRES spectrograph on the 10 m Keck-I telescope. New CCD photometric observations from 1992 to 1996 suggest that V19 was near minimum light at the time each spectrum was acquired, a conclusion that is also supported by spectral analyses using Kurucz model atmospheres. The line widths in the HIRES spectra are quite narrow, with v sin i less than or equal to 18 km s(-1). Given that V19 is the first anomalous Cepheid (AC) to be observed at such high spectroscopic resolution, the rotational velocity is significant in view of the hypothesis that ACs are coalesced binaries. The upper limit on v sin i for V19 appears to contradict this hypothesis, unless sin i is very small. Radial velocities derived from each HIRES spectrogram, v(r) = + 111.0 +/- 0.1 km s(-1) in 1995 August and v(r) = + 110.9 +/- 0.2 km s(-1) in 1996 June, are smaller than would be expected at minimum light based on previous, though much less precise, radial velocity measurements. The lack of a detectable change in v(r) over 311 days, if reinforced by continued monitoring, would also call into question the evolutionary scenario that explains V19 as the result of mass transfer in a close binary system presently composed of an AC plus a white dwarf, unless the orbit is in the plane of the sky. The atmospheric parameters of V19 at the time of the 1995 HIRES observation (phi = 0.51 +/- 0.03) are estimated to have been T-eff = 6450 +/- 50 K, log g = 2.55 +/- 0.15 dex, and microturbulence xi = 2.5 +/- 0.2 km s(-1), in agreement with photometric light-cycle measurements near minimum light. The spectroscopic estimates are based on theoretical line profile fits to the observed wings of the H I Balmer lines H gamma, H beta, and H alpha, together with an investigation of the ionization equilibria of Fe I:Fe II and Mg I:Mg II, taking into account the effects of non-LTE. Adopting these atmospheric parameters and a distance modulus for NGC 5466 of (m - M)(0) = 16.01 +/- 0.10 gives a spectroscopic mass estimate for V19 of 1.66(-0.5)(+0.7) M. which is significantly larger than the mass of the main-sequence turnoff stars in NGC 5466. Equivalent width measurements of 26 Fe II lines give an iron abundance of [Fe/H] = -1.92 +/- 0.05 dex. This value is consistent with previous determinations of the mean metallicity for NGC 5466 from integrated spectra and from lower S/N echelle spectroscopy of individual red giant stars, to within the quoted uncertainties. The absence of the strongest lines of Zn I suggests a possible underabundance of this element. No evidence is found for strong s-process enrichment, with [Ba/H] = -2.45 +/- 0.13 dex from four Ba II lines, consistent with our upper limits [La/H] less than or equal to -1.93, [Ce/H] less than or equal to -1.69, and [Nd/ H] less than or equal to -1.75, based on the marginal detections or absences of the strongest La II, Ce II, and Nd II lines in the spectral regions recorded. Calcium is enhanced with respect to iron, with [Ca/Fe] = +0.22 dex derived from 16 Ca I lines and 56 Fe I lines, under the assumption of LTE; other alpha-elements (e.g., Mg, Si, S, and Ti) are also enhanced relative to iron. In general, the abundance ratios to iron closely match the normal, well-established pattern for metal-poor stars. This result further constrains evolutionary models attempting to explain anomalous Cepheid stars like V19 as the result of coalescence or mass transfer in an evolved binary system; unlike the velocity evidence, this abundance constraint applies regardless of inclination. Finally, the rate of change of V19's pulsation period was investigated using photometry of new CCD frames taken in 1995 and 1996 with the Dominion Astrophysical Observatory 1.8 m telescope and taken in 1996 with the Hale 5 m telescope at Palomar Observatory. This new photometry was combined with unpublished 1992 and 1993 data (courtesy of M. Corwin) and all earlier photometry available from the literature. If the period is assumed to be constant, then the new O-C diagram suggests a revised period of P = 0.821307(d). However, if the O-C data are fitted allowing P to change linearly with time, then the derived period change rate, beta = dP/dt = -0.28(d) +/- 0.19(d) Myr(-1), would imply that V19 is evolving toward hotter temperature, and its pulsation period at the present epoch (1996; JD = 2,450,250) is instead P = 0.18213010(d) + 0.0000003(d). In either case, the residuals in the O-C diagram suggest a possible long-term (P-orb similar to 10,000 day) binary motion. Any such companion would be far too distant to have contributed to the mass of V19 through Roche lobe overflow.
