ON THE MEASUREMENT OF THE LUNAR DISTURBANCE OF GRAVITY. 105 



is it expressly stated that the observer always stands in exactly the same 

 position, although, of course, it is probable that this is the case. It would 

 be interesting, also, to learn whether any precautions have been taken for 

 equalising the temperature of the level itself. To hold the hand in the 

 neighbourhood of a delicate level is sufficient to quite alter the reading. 

 In one of his letters to me M. d'Abbadie also remarks on the slow 

 molecular changes in glass, which render levels untrustworthy for com- 

 parisons at considerable intervals of time. Although we must admire 

 M. Plantamour's indomitable perseverance, it is to be regretted that his 

 mode of observation is by means of levels ; and we are compelled to regard, 

 at least provisionally, these enormous changes of level either as a local 

 phenomenon, or as due to systematic error in his mode of observation. 



In the Report for 1881 we referred to some observations by Admiral 

 Mouchez, made in 1856, on changes of level. A short paper by Admiral 

 Monchez on these observations will be found in the ' Comptes Rendus ' 

 for 1878, vol. 87, p. 665. I now find that the observations were, in 

 fact, discussed by M. Gaillot, in a paper entitled ' Snr la direction de la 

 verticale a I'observatoire de Paris,' at p. 684 of the same volume. The 

 paper consists of the examination of 1,077 determinations of latitude, made 

 between 1856 and 1861, with the Gambey circle. 



M. Gaillot concludes that the variation from year to year is acci- 

 dental, and that the variation of latitude iu the course of the year is 

 represented by 



cX = + 0"-20 sin p60° (^ -_95)-| 

 L 365-25 I* 



where t is the number of days since January 1. 



By a comparison of day and night observations Le concludes that 

 there is no trace of a diurnal variation. On this we may remark that, if 

 the maximum and minimum occur at 6 p.m. and 6 a.m. (which is, roughly 

 speaking, what we found to be the case), then the diurnal oscillation 

 must necessarily disappear by this method of treatment. 



Individual observations ranged from 2"-48 above to 3'''17 below the 

 mean. On this he remarks : — 



' Ceux qui savent combien I'observation du nadir presente parfois de 

 diflBculte dans un observatoire situe au milieu d'une grande ville, .... 

 ceux-la ne trouveront pas ces ecarts exageres, et ne croiront nuUement avoir 

 besoinde faire intervenir une deviation de la verticale pour lesexpliquer.' 



M. Gaillot concludes by remarks adverse to any sensible deviations 

 of the vertical. 



It seems to me, however, that in the passage about the influence of 

 the traflBc of a great town, M. Gaillot begs the whole question by setting 

 down to that disturbing influence all remarkable deviations of the ver- 

 tical. Our observations, and those of many others, are entirely adverse 

 to such a conclusion. 



M. d'Abbadie, in a letter to me, also expresses himself as to the in- 

 conclusiveness of M. Gaillot's discussion. 



He also tells me that M. Tisserand, in his observations of latitude in 

 Japan, found variations amounting to nearly 7"; and when asked ' How he 

 could be so much in error,' answered ' That he was sure of his observa- 

 tions and calculations, but could not explain the cause of such variations.' 



The following further references may perhaps be useful : — Maxwell's 

 paper on the 306-day inequality in the earth's rotation, which was men- 



