Within the past three months ten new variable stars have been discovered at the Harvard Observatory. The variability of three of these stars was suspected elsewhere and has been confirmed by use of improved methods at this observatory.

Three methods are used here to ascertain the variability of a star. The first, and most commonly used, by means of its spectrum, was invented here. There is at the Observatory a collection of about 100,000 photographic plates of different parts of the heavens, made during the last eight years at this Observatory and at the branch observatory in Peru. When a star is suspected of variability a photograph of it is taken and compared with others of this collection and by this means a complete history of the star during the last eight years is established and any variability during this time may be detected.

The second method, also invented here, is by a comparison of two photographic plates taken one after the other. Two plates taken in this way are placed one upon the other and compared with a microscope. In this way any slight variation may be detected.

A large number of observations have also been made with the meridian photometer to determine the variation of light in stars suspected of variability.

The collection of photographic plates at this observatory is the largest in the country, the glass alone weighing ten tons. With these improved methods more variable stars are found here than at any other place.