ON THE DETERMINATION OF GRAVITY AT SEA. 91 



through a lobby and up a companion-way barely wide enough to permit 

 the entrance of a rotund seaman. 



The point that the writer wishes to make is that, apart from the ordi- 

 nary atmospheric changes of pressure which are troublesome enough, there 

 are fluctuations which are due to the relation of the ship's motion to 

 the speed and direction of the wind. An anemometer chart has only 

 to be studied to show how variable these are, so that, even if the ship 

 does not deviate from her course by a hair's breadth, there is ample reason 

 for expecting changes during the time taken for an observation. 



As far as the determination of gravity is concerned these fluctuations 

 would not matter if both the mercury barometer and the comparison 

 instrument, aneroid or hypsometer, possessed the same amount of ' lag,' 

 but this is not the case, and there arises the possibility that the readings 

 of the two instruments do not give truly simultaneous values of the 

 j)ressure. To take an extreme instance — if the aneroid method of 

 estimating gravity were employed, we should obtain a value for the accele- 

 ration due to gra\aty which at 1.55 p.m. w^ould be 0-883 cm./sec.^ higher 

 than the value calculated at 1-46 p.m. As we wish to measure gravity to 

 •005 cm. /sec. ^ we see that on an ocean liner moving through still air the 

 fluctuations may introduce errors many times larger than this amount, 

 unless special precautions are taken to obviate this source of error. 



As the result of these experiments it appears that a new difficulty is 

 added to the already complex problem, of measuring gravity at sia. 

 Such methods as employ a mercury barometer exposed to the atmosj)here 

 (as distinct from the enclosed type) are one degree less satisfactory than 

 was hitherto supposed. In addition to a measurement of the atmospheric 

 jiressure by the height of a column of mercury, some other instrument 

 for measuring it must be employed, and, if their lags are different, any 

 additional fluctuation in the pressures to which the instruments are exposed 

 adversely affects the accuracy of the method. In the method which the 

 Avriter has already tested, whereby the readings of an aneroid are compared 

 with those of a mercury barometer, the two instruments possess very 

 different rates of lag, so that a momentary change of velocity in the 

 wind will certainly affect one instrument more than the other ; moreover, 

 d. comparatively small change in its direction may change the eddy pressure 

 from a jJositive to a negative quantity. Probably the writer in his 

 laboratory, in the refrigerator of the Ascanius or the Morea, protected 

 by two very well-fitting doors from the rest of the ship, and situated in 

 the bowels of the ship, was as free from this disturbing influence as 

 one could possibly be on board ; but, nevertheless, he was not quite 

 secure against a possible fluctuation during the few seconds necessary 

 for taking the aneroid readings. 



Hecker's experiments were, as far as the writer can gather, carried 

 out in passenger cabins which were presumably well above the water- 

 line, and that is a disadvantage. He refers to one cabin as well ventilated, 

 which might after all not be an advantage. In one voyage his work 

 was done in two cabins, in one of which he had his barometer, and in 

 the other his boiling-point apparatus. Without knowing how the cabins 

 were ventilated it is impossible to say whether the results thus obtained 



