thermometric element consisted of
12 balanced mercury thermometers. Its mode of operation is not clear,
but it probably was similar to that of the thermometer developed by Karl
Kreil in Prague about the same time (fig. 4). Dolland's wind force
indicator consisted of a pressure plate counterbalanced by a string of
suspended weights. Altogether, it is not clear that Dolland's instrument
was superior to Hooke's, or that its career was longer.[16]

The 171 years between these two instruments were not lacking in
inventiveness in this field, but even though inventors set the more
modest aim of a self-recording instrument for a single piece of
meteorological data, their brain children were uniformly still-born.
Then, during the period 1840-1850, we see the appearance of a series of
self-registering instruments which were actually used, which were widely
adopted by observatories, and which were superseded by superior
instruments rather than abandoned. This development was undoubtedly a
consequence of the establishment at that time of permanent observatories
under competent scientific direction.

Long experience had demonstrated to the meteorologists of the 1840's
that the principal obstacle to the success of self-registering
instruments was friction. Forbes had indicated that the most urgent need
was for automatic registration of wind data, as the erratic fluctuation
of the wind demanded more frequent observation than any manual system
could accomplish. Two of the British Association's observers produced
separate recording instruments for wind direction and force in the late
1830's, a prompt response which suggests that it was not the idea which
was lacking. One of these instruments--designed by William
Whewell--contained gearing, the friction of which vitiated its utility
as it had that of a number of predecessors. The other, designed by A.
Follet Osler, was free of gearing; it separately recorded wind pressure
and direction on a sheet of paper moved laterally by clockwork. The
pressure element was a spring-loaded pressure plate carried around by
the vane to face the wind. Both this plate and the vane itself were made
to move pencils through linkages of chains and pulleys.[17] Osler's
anemometer (fig. 5) deserves to be called the first successful
self-registering meteorological instrument; it was standard equipment in
British observatories until the latter part of the 19th century when it
was replaced by the cup-anemometer of