Various suggestions have been made, such as that the revolving star is unequally bright over its surface, and shows us now its brighter and now its duller face; that the two pull one another out of shape when they approach most nearly; that they are permanently elliptical, and turn to us first the broad and then the narrow face. Any or all of these may play some part in the variations of the light. But the most hopeful theory takes into account a very strange fact lately established, viz. that the variable is at its brightest not when approaching its fellow but when coming directly toward us, and at its faintest when receding directly from us. This theory is that a dark and bright star are involved in a kind of thin nebulous cloud, and that the bright revolving star has an enveloping atmosphere. As it moves through the cloud this atmosphere is continually brushed back from its advancing face, and therefore we shall see it when coming straight towards us through a less thickness of atmosphere, and it will look brighter than when it is retreating from us. There are grave difficulties in accepting this view, but it receives some support from the case of β Scorpii, in which, as we saw, a double star is suspected, for quite a different reason, of being surrounded by a thin cloud. And the atmosphere supposed to be surrounding the bright star may resemble the “smoky veil” which we know envelops our own sun, and causes a considerable absorption of light, for these variables are solar stars. The resistance of the cloud to the motion of the revolving star ought in time to shorten its period, and some variables have been found to be shortening their periods. Short-period variables are found chiefly in the Milky Way. XI IRREGULAR AND DOUBTFUL VARIABLES Besides these well-marked classes of long-period, short-period, and Algol variables, there are some stars which seem to vary spasmodically, remaining sometimes for months or even for years without any change; and there are others whose variability is suspected but has never been confirmed. Useful work might be done by amateurs in trying to decide the status of these doubtful stars. It must be borne in mind that red stars are notoriously difficult objects. Two observers comparing one at the same time with the same star will often come to opposite conclusions, showing how difficult it is to compare stars which differ much in colour. Whenever possible, red comparison stars should be selected to avoid this uncertainty. The second star on the list below has a very peculiar type of variation, unlike any other except a northern star, R Coronae, and a third recently discovered. The spectra of RY Sagittarii and R Coronae are also peculiar and resemble one another. The following are bright southern variables suspected to vary, or known to vary without recognised laws: SOUTHERN IRREGULAR AND SUSPECTED VARIABLES