The color temperature and Balmer jump inferred from UBV and Stromgren photometric indices, the low gravity and low metallicity derived from spectrum analysis, and weakness of all CH molecular lines combine to suggest that HD 195636 is in an evolutionary state near the transition between the horizontal branch and asymptotic giant branch. The projected equatorial rotational velocity of PID 195636 is V-e sin i=25 km/s, a value at least 2.5 times greater than that expected if known blue horizontal branch axial rotators in globular clusters conserve envelope angular momentum during horizontal branch evolution. Constancy of the radial velocity on time scales of 3, 10, and 90 days indicates that the axial rotation is not due to co-rotation in a short-period binary. Therefore, it seems most plausible that transfer of angular momentum from core to envelope occurred during HB evolution. The Balmer line profiles are peculiar. H alpha is abnormally shallow, as if the core were filled by emission, and higher members of the Balmer series are significantly broader than those of HD 200564, a slightly cooler giant star of similar metallicity. The space velocity of HD 195636 calculated for an assumed RHB/AGB luminosity, M-V=-0.5, is 470 km/s retrograde, a high but not extraordinary value. (C) 1997 American Astronomical Society.

CS 22966-043 is an ultra-short-period pulsating star with high velocity (RV = -266 km s(-1)) discovered during the course of a search for spectroscopic binaries among blue metal-poor field stars, in progress since 1992. With respect to period (0.0374 days), mean color ((B-V) = 0.24), and metal abundance ([Fe/H] approximate to -2.4), it closely resembles the SX Phoenicis stars found among the blue stragglers in NGC 5053. CS 22966-043 also is the primary of a spectroscopic binary with (probable) period of 430 days. Light-travel time across the projected orbit, as large as 0.0037 days, must be added to the times of observation to combine data obtained in different years with minimal phase dispersion. If CS 22966-043 is, indeed, a blue straggler formed by binary interaction as is now generally believed, then it seems most probable that the interaction was one of mass transfer from the present-day secondary during its post-main-sequence evolution rather than merger of a close binary. The latter option would require that this rare field star was, in addition, a member of a primordial triple system.

The blue metal-poor (BMP) star CS 22966-043 is an SX Phoenicis star and the primary of a spectroscopic binary with a provisional orbital period of similar to 430 days. Radial velocity and UBV photometric observations of this star made in 1998 require downward revision of the orbital period to 319 days. The BMP star CS 29499-057 also appears to be an SX Phoenicis star with small amplitude (Delta V similar to 0.04 mag) and short period (P = 0.0417 days), on the basis of photometric and radial velocity observations obtained in 1998. There is some indication that it too may be the primary of a spectroscopic binary. Three other BMP stars have radial velocity standard deviations greater than those of 17 BMP radial velocity standards. We suggest that they may be small-amplitude SX Phoenicis stars. Finally, the BMP star CS 29497-017 is shown to be a short-period velocity variable (P = 0.302 days) on the basis of observations accumulated over an interval of 2200 days, but we were unable to detect a light variation in 1998 July. Therefore, the nature of the velocity variation of this object remains uncertain.

As we enter the third millennium, I reflect on what has been accomplished by the search for extremely metal poor stars during the past 30 years, what might be done better, and what the future holds for this subject.

The radial velocity behavior and chemical compositions of sixty-two blue metal-poor (BMP) stars have been established from more than 1200 echelle spectra obtained at Las Campanas Observatory from 1992 through 1999. Analysis of survey spectra provides abundances for this sample, which we use to calibrate the K line versus B-V relation. Forty-four of the stars have [Fe/H] < - 1, while eighteen lie on -1 < [Fe/H] < 0. One star, the SX Phe variable CS 22966-043, appears to be the most extreme example of a rare abundance class characterized by alpha-element deficiencies, high [Cr/Fe], [Mn/Fe], and [Ti/Fe], and extremely low [Sr/Fe] and [Ba/Fe]. Of the 62 stars, 17 appear to have constant radial velocities, while 42 are definite or probable members of binary systems. The binary fraction of BMP stars appears to be independent of chemical composition.

We have monitored the radial velocities of seven very metal-poor stars with abnormally strong CH G bands. On the basis of their colors four are giants (gCH) and three are subgiants (sgCH). Three of the gCH stars are definite spectroscopic binaries with long periods, 505 days

We report the results of a coherent study of a new class of halo stars defined on the basis of the chemical compositions of three metal-poor objects (Fe/H] similar or equal to -2) that exhibit unusually low abundances of alpha-element (Mg, Si, Ca) and neutron-capture (Sr, Y, Ba) material. Our analyses confirm and expand on earlier reports of atypical alpha- and neutron-capture abundances in BD + 80degrees 245, G4- 36, and CS 22966-043. We also find that the latter two stars exhibit unusual relative abundance enhancements within the iron peak (Cr, Mn, Ni, Zn), along with what may be large abundances of Ga, an element not previously reported as being observed in any metal-poor star. These results provide further evidence that chemical enrichment and star formation histories varied from region to region within the Milky Way halo. Comparing the chemical abundances of the newly identified stellar population to supernova model yields, we derive supernova ratios of Type Ia versus Type II events in the range of 0.6 less than or similar to (N-Ia/N-II)(New Pop) less than or similar to 1.3. For the Sun, we derive 0.18 +/- 0.01 < (N-Ia/N-II)(.) < 0.25 +/- 0.06, supernova ratios in good agreement with values found in the literature. Given the relatively low metallicity and relatively high (N-Ia/N-II) ratios of the low-alpha stars studied here, these objects may have been born from material produced in the yields of the earliest Type Ia supernova events. We also report the results of a preliminary attempt to employ the observed chemical abundances of low-metallicity stars in the identification, and possible cosmic evolution, of Type Ia supernova progenitors, and we discuss the limitations of current model yields.

We have conducted spectrum analyses of 24 field metal-poor ([Fe/H] < -2) red horizontal-branch (RHB) stars identified in the HK objective-prism survey and 6 such stars in the globular cluster M15, based on high-quality spectra (R similar to 40,000, S/N similar to 100) obtained with the Magellan Inamori Kyocera Echelle spectrograph at the Clay 6.5 m telescope at Las Campanas Observatory. The atmospheric parameters of the RHB stars provide interesting bridges between turnoff stars of similar temperature and red giant branch (RGB) stars of similar gravity, and they permit investigations of abundance trends [X/Fe] versus [Fe/H] in a relatively unexplored region of the temperature- gravity plane. We find that the T-eff, log g, v(t), and [Fe/H] values determined from our spectra are consistent with expectations from literature spectroscopic studies of other evolved metal-poor stellar classes. We show that the RHB stars have abundance distributions that are consistent with typical halo stars of similar metallicities. The photometric and spectroscopic gravities of the M15 stars differ by amounts that grow with declining temperature. We use a regression derived from these differences to calculate photometric gravities for the field RHB stars. Then we use the locations of the field RHB stars among the evolutionary tracks of Cassisi et al. in the log g versus log T-eff plane to estimate their masses and lifetimes as RHB stars. We use these lifetimes to estimate the size of the metal-poor HB population from which they arise. Then, using counts of HB and RGB stars in metal-poor globular clusters, we conclude that the number of metal-poor RGB stars at high latitudes (vertical bar b vertical bar > 30 degrees) brighter than V = 15 exceeds those identified in extant objective-prism surveys by more than an order of magnitude. Finally, we deduce the effective temperature of the fundamental red edge of the metal-poor RR Lyrae instability strip, log T-eff(FRE) = 3.80 +/- 0.01, from the interface between the temperature distributions of metal- poor field RHB stars and the RR Lyrae stars of similar [Fe/H] in five metal- poor globular clusters.

We have conducted a photometric and high- resolution spectroscopic analysis of the high- latitude (l = 1 degrees, b = -55 degrees) metal- poor RR Lyrae star TY Gru (= CS 22881-071). We find this star to have large overabundances of carbon and neutron- capture elements. Mass transfer from an asymptotic giant branch binary companion prior to its RR Lyrae evolutionary state is the simplest explanation of this circumstance. Unfortunately, TY Gru is afflicted by the Blazhko effect, which greatly complicates accurate measurement of the motion of its center of mass. We have not yet detected the small orbital motion that would be produced by a distant degenerate companion of TY Gru.

We combine Spitzer and ground-based observations to measure the microlens parallax of OGLE-2005-SMC-001, the first such space-based determination since S. Refsdal proposed the idea in 1966. The parallax measurement yields a projected velocity (v) over tilde similar to 230 km s(-1), the typical value expected for halo lenses, but an order of magnitude smaller than would be expected for lenses lying in the Small Magellanic Cloud itself. The lens is a weak (i.e., non-caustic-crossing) binary, which complicates the analysis considerably but ultimately contributes additional constraints. Using a test proposed by Assef and coworkers, which makes use only of kinematic information about different populations but does not make any assumptions about their respective mass functions, we find that the likelihood ratio is L-halo/L-SMC = 20. Hence, halo lenses are strongly favored, but Small Magellanic Cloud (SMC) lenses are not definitively ruled out. Similar Spitzer observations of additional lenses toward the Magellanic Clouds would clarify the nature of the lens population. The Space Interferometry Mission could make even more constraining measurements.