550 REPORTS ON THE STATE OF SCIENCE. 1916. 



2. The Marine Barometer, 



In the original Report the characteristics and behaviour of this instru- 

 ment are considered with particular reference to the work of Stokes, Chree, 

 and Hecker. It must suffice here to indicate the nature of the discussions. 



1. Construction. — Hecker claims advantages over the Kew pattern 

 for a capiUary constriction with symmetrical funnel-shaped ends and a 

 large space above the mercury. 



2. Lag. — It is clear from Stokes' and Chree's investigations upon the 

 lag of marine barometers at land stations that the viscous resistance to the 

 flow of mercury in the capillary is not the dominant cause of lag. Surface 

 tension effects must be taken into account. 



Chree concludes from land observations that the barometer with the 

 smaller lag possesses a smaller mean error. Hecker, from sea observations, 

 comes to the opposite conclusion. 



Without going so far as Chree in saying that ' at sea the effect of lag 

 upon the average marine barometer is exceedingly small,' the present 

 research favours the view that at sea the lag is less important than on 

 land, probably because the regular throbbing of the engines is more effi- 

 cacious in eliminating unsy mmetrical surface tension effects than per- 

 functory tapping on land. In view of the theoretical uncertainty and the 

 practical difficulty of reading a barometer at sea, it would appear pre- 

 ferable to place the barometer and the other apparatus with which its 

 readings are to be compared in a chamber in which the rate of change of 

 pressure can be controlled and measured and reduced to a small and 

 determinate quantity, if not to zero, rather than to trust to measurements 

 of the variations of the atmospheric pressure, which, since the ship is 

 moving, are likely to be more rapid even than those encountered by a 

 fixed barometer at a land station, and which are seldom linear for any 

 considerable period. 



It is suggested that the chamber should be large enough to contain 

 photographically recording aneroid and mercmy barometers and furnished 

 with an auxiliary aneroid which could be used as a regulator ; by means 

 of an electric contact operating a relay working a rotatory pump it should 

 be possible to maintain a nearly constant pressure. 



The error introduced by fluctuating pressure is shown in fig. 14 for a 

 harbour station ; compared with other consecutive Morea deviations at sea, 

 fig. 9, the deviations are large. 



3. Pumping. — A vertical acceleration of the barometer may be occa- 

 sioned by the rise or fall of the ship as a whole, or by rolling and pitching 

 about a longitudinal and transverse axis respectively, if the apparatus is 

 not in the centre of the ship. The vertical motion adds an acceleration 

 to that due to gravitational attraction, and the problem is complicated by 

 the fact that this dynamic acceleration may not be symmetrical. 



The constriction is introduced for the purpose of freeing the static 

 attraction from the dynamic acceleration, but though it reduces it does 

 not eliminate the pumping. The damping is always such that the free 

 vibration of the mercmy is aperiodic. 



When the mercury is pumping it is necessary to take the mean of 

 successive maxima and minima readings. The practice of reading only 

 the highest (or lowest) point, which at one time received official sanction, 

 is deprecated. It is difficult to set and read the barometer and to record 

 the reading iu the half period of the wave, but with a dial form of instru- 



