The field covered by this report overlaps particularly with the work of Commissions 29 and 45 whose reports should be consulted. I have been greatly helped by reports on Soviet work by Professor Kukarkin and Dr Boyarchuk and a bibliography of European literature by Bloch.

The Sonneberg programme for determining spectral types of variables has been described by Götz and Wenzel (AN, 290, 99) and several lists have been published in MVS. Bakos is studying high dispersion spectra of a number of variables that are members of visual binary systems.

Stephenson and Terrill (ApJ, 147, 148) and Houk (IB, 228) have given spectral types for a considerable number of long period variables. Terrill (AJ, 74, 413) has made an extensive study of the variation of spectrum with phase for 51 red variables. Deutsch, Keenan and Garrison have continued their extensive programme of spectral classification of Mira variables. Keenan and Dessy (ApJ, 146, 583) suggest that the relatively early spectral type of the very long period Miras (R Cen, R Nor etc.) showing double maxima indicates that the fundamental parameter is the half-period (between successive maxima). A similar conclusion (ApJ, 93, 380; MN, 125, 367) was reached earlier by others on the basis of the correlation with period of the difference between absorption and emission velocities. Feast has used K subgiant companion of U Men to derive an absolute magnitude of — 0m.5 for the Mira star. The radial velocity of the companion agrees with the absorption velocity of the variable near maximum (Obs.).