Curves of Growth and Line Contours.

The curve of growth and the line contours of ~ Canis Majoris (cF8) are explained by assuming large- scale turbulence with = 22 km/sec or-less probably-rotation with an equatorial velocity v sin i 30 km/sec. Moreover, the intensity of the continuous background in the range 4000-4100 A may be depressed by the crowding of lines to 70 per cent of its original intensity. It is also possible that on top of a normal reversing layer having a turbulent velocity of the order of 5-10 km/sec there is an optically thin layer produced by a field of interlacing prominences which may be described by a large turbulent velocity. The differences which K. 0. Wright found between the curves of growth for neutral and those for ionized atoms, as well as low-and high-excitation states, in solar-type stars can be interpreted by as- suming that the turbulent velocity ~ increases with decreasing pressure, following the law of continuity. The remaining inconsistencies are traced to a probable "coupling" in the simultaneous determination of turbulent velocity and excitation temperature. Reasons are given for the conclusion that the solar excita- tion temperatures are closer to 5400° or 5700° than to 4800° K. The use of "solar transition probabilities" is considered critically. The meaning of effective turbulent velocities is discussed. Errors resulting from the use of a single curve of growth for an entire stellar spectrum are estimated not to exceed z~ log NH = ± 0.47 if suitable mean values are used. The aerodynamic aspects of turbulence in stellar atmospheres are reviewed, following recent developments of the theory of turbulence