engaged his attention. Geology was the first subject that he took up seriously. He became a fellow of the Geological Society in 1870, and at the British Association meeting of 1878 he read a paper on the filtration of water through triassic sandstone. Between 1882 and 1889 he made an elaborate series of experiments on the movement of underground water as affected by barometric and lunar changes. A paper on a different subject, 'the determination of the vertical and lateral pressures of granular substances,' which appeared in the 'Proceedings of the Royal Society' for 31 Jan. 1884, embodied the results of elaborate experiments made for the purpose of furnishing data to engineers and builders of storehouses.
Meanwhile his attention had been turned to astronomical observation. In 1878 he had a 7-inch refractor by Cooke at his home at Rock Ferry, Birkenhead, which he used for visual observation, but a few years later he applied himself with zeal to the advancing practice of stellar photography. In 1883, a year after his removal to Kennessee, Maghull, near Liverpool, he experimented in photographing stars with ordinary portrait lenses varying in aperture between three-eighths of an inch and five inches. After consideration of the results of these experiments and comparisons with the photograph of the nebula in Orion by {{DNB lkpl|year=12|Common, Andrew Ainslie|Andrew Ainslie Common [q. v. Suppl. II], he ordered from Grubb of Dublin a 20-inch silver-on-glass reflector of 100 inches focal length, the photographs to be taken directly in the focus of the mirror to obviate any loss of light by a second reflection, and the photographic telescope to be mounted on the same declination axis as the 7-inch refractor, one being the counterpoise of the other (Monthly Notices R.A.S. xlvi. 99).
At the meeting of the Royal Astronomical Society of January 1886, Roberts, who was at the time the president of the local Astronomical Society at Liverpool, reported taking during the past year 200 photographs of stars which might be measured for position, as will as long exposure photographs of the Orion nebula, the Andromeda nebula, and the Pleiades. At the November meeting in the same year he presented a photograph of the Pleiades taken with his 20-inch reflector with exposure of three hours, which showed the stars Alcyone, Maia, Merope, and Electra surrounded by nebulosity extending in streamers and fleecy masses till it seemed almost to fill the spaces between the stars and extend far beyond them. This photograph was accepted as revealing structure about the group never before seen or suspected. A photograph of the great nebula in Andromeda presented at the meeting of December 1888, which suggested that the object is of the spiral type, evoked considerable interest because it was supposed to illustrate the main idea of the nebular hypothesis. Photographs of the great nebula in Orion, presented a few months later, were equally successful. Roberts persistently urged the superiority of the reflector over the refracting telescope, a view which has since received much confirmation. In the early years of his work Roberts designed an instrument, the pantograver, an example of which was made for him by Mr. Hilger, for transferring mechanically the images on a photographic negative to a copper plate, to be used for making reproductions (Monthly Notices, Nov. 1888).
Roberts attended by invitation the Conference of Astronomers at Paris in 1887 which initiated the international survey of the heavens by photography, but took no part in the scheme, which was entrusted to professional astronomers at national observatories with instruments of a uniform type. In order to continue his work on the nebulae and star clusters in a clearer atmosphere than that of Liverpool, he finally settled in 1890 at Crowborough Hill, Sussex, in a house appropriately named Starfield. There Mr. W. S. Franks, an astronomer and skilful photographer, became his working assistant, and Roberts confined himself to organisation and supervision. Month by month for several years he exhibited at the Royal Astronomical Society splendid photographs of remarkable objects in the sky taken with his reflector. Two volumes of selections of Roberts's photographs of stars, star clusters, and nebulae, 125 reproductions in all, appeared respectively in 1893 and 1899. In 1896 Roberts, following the example of Professor Barnard in America, added to the equipment of his observatory cameras with portrait lenses of different types, in order to compare their photographic results with those of the reflecting telescope (cf. a discussion on the relative efficiency of the two methods between Roberts and Professor Barnard in R.A.S. Monthly Notices, lvi. 372, lvii. 10, lviii. 392). Between 1896 and 1902 Roberts prepared photographs of fifty-two regions of the sky called 'nebulous' by Sir William Herschel, made with his reflector and with a portrait lens of 5 inches aperture made by Messrs. Cooke of York. No diffused nebulosity was shown on forty-eight of
