320 Kew Observatory. 



observer give a satisfactory definition of that point ? Was 

 either aware of the gradual change which takes place in a 

 thermometer by age ; or in graduating an instrument, was any 

 allowance made for the unequal diameter of the bore at different 

 parts of the same tube ? These and other questions might 

 well be asked ; nor do we err in stating that errors in baro- 

 meters of that period might often be reckoned by tenths of an 

 inch, errors of thermometers by degrees. 



But day was now beginning to dawn, the public were 

 gradually becoming aware of the practical importance of 

 meteorology ; the laws of storms (for even storms have laws) 

 were more observed, and while Admiral Fitzroy applied himself 

 to the task of foretelling weather, the Kew committee set 

 themselves to that of improving instruments ; for in the peace- 

 ful as well as in the warlike arts, one man furbishes the weapon 

 which another man wields. 



It was at this stage that the committee were fortunate in 

 securing the valuable co-operation of the late Mr. Welsh as 

 superintendent of the observator}^. One of their first acts 

 was to recommend a pattern for barometers to be used at sea, 

 and instruments after their model have since been very exten- 

 sively employed by Admiral Fitzroy in the department under 

 his control. 



Another important point was to obtain at Kew the means of 

 readily determining the errors of meteorological instruments, pre- 

 vious to which it was essential to construct an accurate standard 

 barometer, to which all others might bo referred. Let not our 

 readers imagine that this was an easy task, for in order to avoid 

 the influence of capillarity, it was necessary that " the internal 

 bore of the tube to be filled with mercury should bo at least 

 one inch in diameter. This, after much preliminary difficulty, 

 Mr. Welsh accomplished, by a method which obviated the 

 trouble of boiling the mercury in the tube — in all cases a dif- 

 ficult operation, but with a tube of such a bore nearly im- 

 possible. 



I Laving procured their standard of reference, something'- more 



however, wanting before barometers could be properly 



tested; BO doubt, by suspending instruments in the same room 



with (lie standard, the errors of these might bo obtained, but 



only for the existing atmospheric pressure, whatever that might 



happen to be at the time of comparison. Hut for marine 

 barometers, with no cistern adjustment, it was essential to 

 know the error at various points, and clearly it would not do 

 to wait for a storm in order to oompare together instruments 

 at a low pressure, or for exeept ionably line weather, in order 

 to Compare them when ihe pressure was high. 



Evidently the only plan was to obtain the means of pro- 



