[Illustration: Figure 16.--Richard's registering aneroid barometer, an instrument used at the U.S. Weather Bureau about 1888. The Richard registering thermometer is similar, the aneroid being replaced by an alcohol-filled Bourdon tube. (_USNM 252981; Smithsonian photo 46740-C_.)] To attribute the success of self-registering instruments in the late 19th century to the unquestionable improvements in the techniques of the instrument-maker is to beg the question, for it is by no means clear that the techniques of the 17th-century instrument-maker were unequal to the task. It should also be noted that the photographic and electromagnetic systems of the 19th century seem to have been something of an interlude, for some of the latest and most durable (all of Draper's and Richard's instruments and Marvin's barograph) were purely mechanical instruments, as had been those of Hooke and Wren. If we conclude that the 19th-century instruments were more accurate, we should also recall Forbes' comments upon the question of instrumental accuracy. What, then, was the essential difference between the 17th and 19th centuries that made possible the development of the self-registering observatory? It would appear to have been a difference of degree--the maturation in the 19th century of certain features of the 17th. The most important of these features were the spread throughout the western world of the spirit that had animated the scientific societies of Florence and London, the continued popularity of the astronomical observatory as an object of the philanthropy of an affluent society, and the continued existence of the nonspecialized scientist. Under these circumstances such nonmeteorologists as Wheatstone, Henry, Hough, Wild, and Secci had the temerity to range over the whole of the not yet compartmented branches of science and technology, fully confident that they were capable of finding thereby a solution to any problem important enough to warrant their attention. FOOTNOTES: [1] On early meteorological instruments see A. Wolf, _A History of Science, Technology and Philosophy in the Sixteenth and Seventeenth Centuries_, New York, 1935, and E. Gerland and F. Traumüller, _Geschichte der physikalischen Experimentierkunst_, Leipzig, 1899. On the recognition of the meteorological significance of the barometer by Torricelli and its meteorological use in 1649 see K. Schneider-Carius, _Wetterkunde Wetterforschung_, Freiburg and Munich, 1955, pp. 62, 71. [2] Bacon's book emphasizes "direct" and "indirect" experiments, and calls for the systematization of observation, but it does not mention instruments. It is reprinted in Basil Montagu's _The Works of Francis Bacon, Lord Chancellor of England,_ London, 1825, vols. 10 and 14. [3] Wolf, _op. cit._ (footnote 1), pp. 312, 316-320. The interest of the Royal Society in the barometer seems to have been initiated by Descartes' theory that the instrument's variation was caused by the pressure of the moon.