the furtherance of meteorology one of its first objectives.

The Kew Observatory set a pattern for systematic observation in England
as, from 1855, did the Smithsonian Institution in the United States. The
instruments used differed little from those in use at Mannheim over half
a century earlier[6] (fig. 1). They were undoubtedly more accurate, but
this should not be overstressed. Forbes had noted in his report of 1832
that some scientists were then calling for a return to Torricelli, for
the construction of a temporary barometer on the site in preference to
reliance on the then existing manufactured instruments.

The First Self-Registering Instruments

From the middle of the 17th century meteorological observations were
recorded in manuscript books known as "registers," many of which were
published in the early scientific journals. The most effective
utilization of these observations was in the compilation of the history
of particular storms, but where a larger synthesis was concerned they
tended, as Forbes has shown, to show themselves unsystematic and
non-comparable. The principal problems of meteorological observation
have been from the outset the construction of precisely comparable
instruments and their use to produce comparable records. The former
problem has been frequently discussed, and perhaps, as Forbes suggests,
overemphasized. It is the latter problem with which we are here
concerned.

The idea of mechanizing the process of observation, not yet accomplished
in Forbes' time, had been put forward within a little over a decade of
the first use of the thermometer and barometer in meteorology. On
December 9, 1663, Christopher Wren presented the Royal Society with a
design for a "weather clock," of which a drawing is extant.[7] This
drawing (fig. 2) shows an ordinary clock to which is attached a
pencil-carrying rack, geared to the hour pinion. A discussion of the
clock's "reduction to practice" began the involvement of Robert Hooke,
who was "instructed" in September 1664 to make "a pendulum clock
applicable to the observing of the changes in the weather."[8] This
tribute to Hooke's reputation--and to the versatility of the mechanic
arts at this time--was slightly overoptimistic, as 15 years ensued
before the clock made its appearance.

[Illustration: Figure 2.--A contemporary drawing of Wren's "weather
clock." (Photo courtesy Royal Society of London.)]

References to this clock are frequent in the