Self-recording barometers and thermometers were more vulnerable to the influence of friction than were wind instruments, but fortunately pressure and temperature were also less subject to sudden fluctuation, and so self-registration was less necessary. Nevertheless, two events occurred in the 1840's which led to the development of self-registering instruments. One event was the development of the geomagnetic observatory, which used the magnetometer, an instrument as delicate as the barometer and thermometer, and (as it then seemed), as subject to fluctuation as the wind vane. The other event was the development of photography, making possible a recording method free of friction. In 1845 Francis Ronalds at Kew Observatory and Charles Brooke at Greenwich undertook to develop apparatus to register the magnetometer, electrometer, thermometer, and barometer by photography.[18] This was six years after Daguerre's discovery of the photographic process. The magnetometers of both investigators were put into use in 1847, and the barometers and thermometers shortly after. They were based on the deflection--by a mirror in the case of the magnetometer and electrometer and by the mercury in the barometer and thermometer--of a beam of light directed against a photographic plate. Brooke exhibited his instruments at the Great Exhibition of 1850, and they subsequently became items of commerce and standard appurtenances of the major observatory until nearly the end of the century (fig. 6). Their advantages in accuracy were finally insufficient to offset the inconvenience to which a photographic instrument was subject. Before 1850 the British observatories at Kew and Greenwich (the latter an astronomical observatory with auxiliary meteorological activity) had self-registering apparatus in use for most of the elements observed. Self-Registering Systems In 1870 the Signal Corps, U.S. Army, took on the burden of official meteorology in the United States as the result of a joint resolution of the Congress and in accordance with Joseph Henry's dictum that the Smithsonian Institution should not become the permanent agency for such scientific work once its permanency had been decided upon. Smithsonian meteorology had not involved self-recording instruments, and neither did that of the Signal Corps at the outset "because of the expense of the apparatus, and because nothing of that kind was at that time manufactured in this country."[19] But almost immediately after 1870 the Signal Corps undertook an evidently well-financed program for the introduction of self-registration. "Complete outfits" were purchased, representing Wild's system, the Kew system as made by Beckley, Hipp's system (fig. 8), Secci's meteorograph (figs. 9, 10), Draper's system, and Hough's printing barograph and thermograph. Of these only the Kew system, the photographic system already mentioned, could have been obtained before 1867. [Illustration: Scale about 1-16th.