Very curious are the systems thus discovered. Algol stars are extraordinarily light for their size, their density being always less, and sometimes immensely less, than that of water, and the companions are usually extraordinarily close together. In some pairs they seem to be actually touching. Nearly all are Sirian stars; a few are of Orion and solar types. Sometimes the companion star gives light also, instead of being dark, and then we have a different type of variation. There are two eclipses in one revolution, each star passing alternately behind the other, but neither is a very dark eclipse, only a lessening of light. β Lyrae was the first star of this kind to be discovered; its southern counterpart is U Scuti, which varies from magnitude 9.1 to 9.6, and runs through its two maxima and two minima in less than twenty-three hours! It is a Sirian star. X SHORT-PERIOD VARIABLE STARS Another extremely interesting class of variable stars runs through the periods as punctually as the Algol stars, but the light is varying up and down the whole time without any period of steadiness. The length, too, is often much greater, though not nearly so great as that of the Mira variables. Here are a few stars of this class bright enough to be observed with a binocular, at least at maximum: SOUTHERN SHORT-PERIOD VARIABLES +----------------------+----------+----------+-------+ | Name. | Maximum | Minimum |Period.| | |Magnitude.|Magnitude.| | +----------------------+----------+----------+-------+ | | | | Days. | |W Sagittarii | 4.3 | 5.1 | 7½ | |κ Pavonis | 4.3 | 5.2 | 9 | |_l_ Carinae | 3.6 | 5.0 | 35½ | |S Crucis | 6.5 | 7.6 | 4½ | |V Centaurii | 6.4 | 7.8 | 5½ | |S Trianguli australi | 6.4 | 7.4 | 6½ | |S Normae | 6.6 | 7.6 | 9¾ | +----------------------+----------+----------+-------+ The three first on the list have been discovered to be spectroscopic binaries, the motions varying with the light, and as this is found to be the case with all those whose motions are known, there can be little doubt that all variables of this class are binaries. Their orbits, however, are somewhat larger than those of the Algol variables, and are tilted towards us so that neither star can be seen to pass in front of the other. For the gradual and continual change of light cannot possibly be caused by an eclipse. Yet it seems clear that it is in some way connected with the revolution of two stars about one another.