meteorological literature, and appears to have been defective or out of
fashion with its time, which was concerned with the introduction of
photographic instruments. Wheatstone's work was rediscovered, along with
that of several other much earlier inventors, by the determined
observatory directors of the 1860's.

Of the five systems developed at that time, four used electromagnetic
registration, only Draper adhering to a mechanical system (see fig. 11).
For temperature measurement Secci and Hough used Wheatstone's electrical
system with a mercurial thermometer (fig. 12), but the other four
utilized a physical principle which had been proposed periodically for
at least a century--the unequal thermal expansion of a bimetallic strip.
This principle had been utilized by watchmakers for a quite different
purpose--the temperature compensation of the watch pendulum--but its
possibilities as a thermometer had been known long before the mid-19th
century.[27]

[Illustration: Figure 8.--Hipp's registering aneroid barometer, with a
telegraphic printer. (_USNM 314544; Smithsonian photo 46740-D._)]

For the measurement of pressure, Secci, Wild, and Draper adopted, or
rediscovered, the balance barometer devised by Wren in the 17th century.
In this type of instrument (see figs. 13, 15) either the tube or the
reservoir of the barometer is attached to one arm of a balance, the
equilibrium of which is disturbed by the movement of the mercury in the
instrument.[28]

[Illustration: Figure 9.--Front and rear views of Secci's meteorograph,
1867. (From Lacroix, _op. cit._ footnote 22.)]

Hough's barometer was an adaptation of the electrical contact
thermometer. The movement of the mercury over a certain minute distance
within the tube served as a switch to energize an electrical recording
system. Hipp, who was perhaps the latest of this group, first applied
the aneroid barometer (fig. 8) to self-registration. The idea of the
aneroid--an air-tight bellows against which the atmospheric pressure
would act--had been advanced by Leibniz in the 17th century and had been
the subject of a few abortive experiments in the 18th century. Not until
1848 was an instrument produced that was acceptable to users of the
barometer.[29]

As a wind velocity instrument all six systems used the cup-anemometer
developed by Robinson in 1846, an instrument whose chief virtue was the
care which its inventor had taken to work out the relationship between
its move