a SII 1194 SII 1201 CI 1175 90+ (3 lines) : (2 lines} {5 lines) 80} T TT T T ‘ i T 1 “a 100 4 . 70} : 1206.52 60; 50r Theoretical . from model 4 40+ spectrum 30+ 7 approximate type B15¥V oot iO; 1180 1170 1160 1150 1220 1210 1200 1190 1240(A) 1230 Fig. 3. Predicted strong stellar lines in the neighborhood of 1200 A. those of Mg II at 2800 A. Calculations suggest that these lines will be quite strong in B-type spectra although less in the outermost layers of the models. Thus study of the observed profiles and displacements of the strong resonance lines should yield information about the electron temperature and pressure and the velocity field in the outermost parts of the stellar atmosphere, especially whene an improved theory of line-formation is used. Mor- strong than the resonance lines of C IV and Si IV. This region contains a few strong lines of Si III and C III and very many lines of the second and third spectra of the metals. Estimates of the strength and profiles of the lat- pear in Fig. 4. The lines from the third spectra of the metals should be easily observed, with a spectral purity of the order of 0.5 to 1 A. The line profiles shown in Fig. 4 have been calculated with a so-called microturbulence equal to three times the thermal Doppler broadening. Observation has shown that the spectral lines of the standard “sharp line” stars are not really sharp, and that predicted profiles most easily can be madetofit the observed profiles by introduction of the arbitrary parameter microturbulence. The adopted motions are small, the root-mean-square velocity of an atom or ion being about 12 kilometers per second instead of 3 kilometers per second as is characteristic of heavier atoms at the temperatures in B1.5 atmospheres. There are many ways in which such a velocity distribution could arise in a stellar atmosphere. Trial computations have shown that the adopted velocity field, which is used in descrip- tion of the shape of the line-absorplines of the second spectra will be tion coefficient, has far more influence weak, but that the lines of the third tinuous spectrum from O and stars on the shape and strength of lines (Fig. 1) at wavelengths between 911.6 spectra of the metals will be deep and predicted at wavelengths longer than and 1500 A; rather the spectrum of as wide as 1 A. The spectra of B 2000 A than have differences in the stars between 1900 and 3000 A are temperature-pressure structure of the main-sequence stars will be cut up by many deep absorption lines—some of expected to look rather like spectra - models due to the inclusion or omiswhich are shown schematically in -Of F stars in the spectral region obsion of line blanketing in the process Fig. 2. served with ground-based instruments; of construction of models. many weak lines and a moderate numThe wavelength region between 1900 Observation and interpretation of the and 3000 A has been studied at Utrecht. ber of moderately strong, deep lines line spectra of B stars in the wavelength The only resonance lines expected are will be present; typical line profiles apregion between 1900 and 3000 A should lead to new and improved estimates of the abundances of the metals (Fe, Mn, Cr, Ti) in the atmospheres -05 +05 -05 +05 -05 +05 -05 +05 A re d tT T | OD t T T i er i ] I of B stars. These estimates should be 100 —~ 100 quite reliable because the line spec90+ \ + 90 trum in the region between 1900 and 3000 A is formed through exactly the sol | 80 same parts of the atmosphere as is 70 Cri Fell Fell NiI | P 283563 275329 2739.54 221648 70 the part of the spectrum observed with ground-based instruments. (This state60~ 60 Model B13 ment appears obviously true from the 50+ B15 50 relative value of the continuous-absorption coefficient in the region between 100 100 1900 and 6800 A.) Thus good models 90 | | -| 90 for the relevant layers of the atmo80}. + BO sphere can be established and controlled by normal ground-based obser7Or | 70 Til Cri MnIl Fell vation. 6oL 2516.01 211753 _| 2068.96 2103.80 | 6, The reason the abundances of the metals in B stars cannot be estimated 50+ + 50 easily from the parts of the spectrum 40 4 40 that are observed by ground-based instruments is that too few lines of suf30+ + 30 ter lines (5, 6) have shown that the ton’s calculations show that it is quite unreal to expect observation of a con- 1 -05 { t +05 ! ! -05 al t +05 ! I -05' ! t +05 l ft -05 i +05 A Fig. 4. Predicted lines from the second and third spectra of the metals in the spectral region 1900 to 3000 A. 1276 ficient strength occur tn the accessible spectral region to permit successful applications of the usual methods of spectrum analysis. It will be interesting to SCIENCE, VOL. [58