1 72 Proceedings of the Royal Society of Victoria. 



behaves in a iiKist eccentric manner, and tliat its variations 

 are but very slif^-htly indicated by the air pump test made at 

 the Observatory on 20th December, 1894. The \\ inch aneroid 

 worked very mucli better, l)ut did not behave as it did under 

 the air pump, whilst the 2^^ inch aneroid shows very well with 

 the exception of one discordant observation, and its l^ehaviour 

 is very similar to what it was under the air pump. 



Taking the whole of the results they fully justify (so far as 

 they go) Mr. Edward Whymper's conclusion, " that the test which 

 is usually applied of comparing for brief periods (minutes or 

 hours) aneroids with i>-iov.r..i..;-i i 



.- ^..^ oijj«.ii uure Jiountain niercuiial is at tirst 

 sight very peculiar, but is, undoubtedly, due to varying capillary 

 action in the small instrument. No measurements of the heights 

 of the meniscus were taken, but it was apparent that these were 

 continually varying both in upper and lower limbs though 

 principally in the latter where the meniscus at times entirely 

 disappeared, and at other times exceeded considerably that in the 

 upper limb. In every case the instrument was well tapped prior 

 to taking a reading. It would appear that the readings of this 

 barometer could be depended on to about 0-03 inch, and the 

 eri'or would be independent of altitude. 



The boiling-point experiments resolve themselves into two 

 sections, one taken with a glass spirit lamp, which was not 

 sufficiently powerful to maintain a good supply of steam, and the 

 other with a brass lamp which generated steam with ease. As 

 might be expected the boiling pcjints given l^y the former are all 

 lower than those given by the latter. One of the observations 

 (that on the 2r)th August, 4 p.m.) should undoubtedly be rejected, 

 as the apparatus was at the time undergoing alterations and 

 consequently the bulb was exposed to a mixture of air and steam 

 instead of pure steam. Taking the second set of observations 

 (sixty-seven in nundjer), the maximum dilierence in the correction 

 to standard is 0*048 inch as against 0"0G8 inch with the 

 Mountain mercurial. 



