Considerations on the Problem of Latitude Variation. 25 



much increase as to account for the whole closing sum. The second 

 and third are nothing more than conjectures. 



Under these circumstances, it does not seem utterly super- 

 fluous to make new suggestions, to be further discussed and inves- 

 tigated by a wider circle. With tliis intention, I venture to declare 

 that the above puzzling subjects may possibly l)e explained l)y the 

 phenomenon above discussed, in connection with the defect of the 

 spirit level. 



Now, provided that the considerations in the earlier part of 

 this essay on the disturbance of the zenith telescope are applicable 

 to tlie case when the Talcott-Horrel)OW observation is made, in 

 which the proximity of the observer is of short duration, the first 

 part of the disturbing efïect only may come into play, so tliat the 

 change of inclination of the telescope, due to the thermal disturb- 

 ance, would not appear in the position of the level bubbles. 

 Therefore, the level bubbles Avill be situated too far north of their 

 due position. The result of this is that the corrections depending 

 on the level reading are positively too large or negatively too 

 small, giving finally too large a value of latitude. The principal 

 part of this error varies inversely with the temperature. So the 

 correction to be applied to the latitude is of negative sign and of 

 varying magnitude, depending on the seasons and the hour of the 

 day. 



Let the star groups selected for the chain method be from I to 

 XII, as is adopted in the International Latitude Service. And 

 further distinguish these groups by suffixes according to coml)ina- 

 tions ; sufiix 1 when coml)ined with the ])receding group and 

 suffix 2 when combined with the following group. The period of 

 observation and the corresponding group are as follows, — 



