STELLAR SPECTRA We deal rather briefly with this branch of the subject since some of the more interesting developments are treated under ASTROPHYSICS, STELLAR EVOLUTION, SUN. By the appearance of their spectra (see SPECTROSCOPY) stars are arranged in a se quence of types 0, B, A, F, G, K, M. This turns out to be the order of decreasing surface temperature. The "effective" sur face temperatures of stars are calculated by Planck's law of radia tion (for theory, see RADIATION) from the relative intensity of the continuous spectrum for different wave-lengths; the higher the temperature the more the region of maximum intensity is dis placed to the violet. For the sun this method of finding the equivalent black-body temperature is checked by calculating also the temperature corresponding to the absolute intensity of its radiation ; so presumably it will apply to other stars, although we have not usually any direct check., Good indirect confirma tion of these temperatures has been afforded by studying the spectral indications of ionization of the elements. (See below.) The following temperature scale is a combination from various sources ; the results are in degrees absolute.
This refers to giant stars; dwarf stars of the same type are about hotter. The radiating power of the surface in heat units is proportional to the fourth power of the temperature.
In all the later types the stars fall into two groups, the giant or diffuse stars, and the dwarf or dense stars. They are most widely distinct in type M, and converge to coalescence in types A and B. If, instead of arranging the stars in order of surface temperature, we put them in order of increasing density, we must start with the giants of type M, follow the giant branch through K, G, F until it coalesces with the dwarf branch in types A and B, and then continue along the dwarf branch in the reverse order A, F, G, K, M. It is natural to suppose that the order of in creasing density is that of evolution, and accordingly we seek a theory of evolution which will be consistent with this order.
Following this supposed order of evolution, the temperature of the surface starts very low, increases to a maximum, and then decreases again, the giant stage being one of ascending tempera ture, and the dwarf stage of descending temperature. This fact, which was stressed in early theories, has probably no great sig nificance, since the surface temperature does not correspond at all closely to the internal temperature. It is more important to note that the mass (corresponding closely with the luminosity) is nearly steady in the giant sequence, perhaps decreasing slightly, and then diminishes rapidly as we pass down the dwarf sequence.
It seems impossible to retain our general ideas of stellar evolution without admitting that the mass of a star decreases substantially during its life-time. Radiation is the principal cause of loss of mass by a star, and hence we are inclined to believe that a large part of the mass of a star is convertible into radiation—in short, matter must disappear, its energy of constitution being released as aether waves. Admitting this source of energy, the age of the stars is of the order of a billion (io") years or more. In fact, we can calculate the time taken for a star to evolve from one type to another by reckoning how long it would take for its observed radiation to carry away the necessary amount of mass. In the latest dwarf stages, when the radiation is feeble, evolution will be very slow, and the sun may continue to shine with increasing feebleness for 30o billion years before it reaches the last stage of the sequence type M (dwarf).
After the first use of spectroscopy to identify chemical ele ments in the atmosphere of a star, the most important advance was the discrimination of "enhanced lines," i.e., lines strengthened in the spark spectrum as compared with the arc. Much of the early progress in this direction was due to J. N. Lockyer and A. Fowler. In general, enhanced lines are those due to the atom with one electron missing (singly ionized atoms), the arc lines being due to neutral atoms. The spectra of all the important ele ments have now been unravelled, and we can say at once which lines belong to neutral atoms, which to singly ionized, which to doubly Ionized atoms, etc. Thus it is found that in some stars the atoms of calcium in a position to form an absorption spectrum are neutral, in others they are singly ionized, in others doubly ionized. Clearly this is a clue to the physical conditions in the outer layers of a star. The modern thermodynamical theory of ionization was brought to bear on the astronomical observations by M. N. Saha in 192o, and has dominated all recent progress. The degree of ionization depends on both the temperature and pressure, the former representing the factor which causes the breaking away of electrons and the latter the factor which de termines their opportunity of recombining. By connecting to gether the data for different elements and controlling the results by reference to other methods of determining stellar tempera tures, much information has been obtained as to the temperature and pressure in the reversing layer. In particular it is found that the pressure is generally between and atmospheres— much lower than used to be supposed.