14 Art. 3-K. Sotome: 



to be noticed tliat the rate of the tilting motion increases from 

 evening until miclniglit. 



In this way we are led to the conclusion that the observed 

 phenomenon is to be understood as principally due to the second 

 and third of the above causes, whose efïects can be looked upon as 

 generally varying inversely with the temperature during the time 

 of our observations. 



The above cited singular Ijehaviour of the zenith telescope was 

 deduced from the regular shift of the level bubbles, as experi- 

 enced by myself. In order to ascertain whether the same pheno- 

 menon is noticed in other ol)servatories, I made enquiry of several 

 latitude observers in \'arious i)arts of the world respecting this sub- 

 ject. Fortunately, almost all the specialists were kind enougli to 

 favour me v/ith detailed answers, for which the writer tenders his 

 most sincere thanks. 



Prof. H. Kimura and Dr. M. Hashimoto of the Mizusawa 

 Latitude (Jbservatory recognized the phenomenon in the zenith 

 telescope there used. We can also see the tendency from the re- 

 sults of their elongation observations, as given in " liesultate des 

 Internationalen lîreitcndienstes," Bd. I., pp. 19-20. In the deter- 

 mination of the value of tlie micrometer during the period 1899 

 Dec. 27-1900 March 11, eastern elongations constantly give larger 

 values than western. This corresponds to the case Avhen the re- 

 duction was made on assumption II. (siqira). 



I could hear nothing from Tschardui observers. But the series 

 of determinations given in "Resultate," Bd. IV., page 74, makes 

 me suppose that there is also the same tendency there. In the 

 observations during the period IDOG January 19 — February 28, the 

 mean value from western elongations (60. "2G8) is decidedly larger 

 than that from eastern elongations (60. "148). This may be caused 

 l)y tlie method of reduction on assumption I. If such is really the 

 the case at Tschardui, we can fairly well account for the fact that the 

 temperature coefficient of the Tschardui micrometer comes out 

 notably lai^ger than those of the other stations. This is because 

 winter observations are principall}" of western elongation while 

 summer observations are mostly of eastern elongation; so that the 



