T907.] - OGBURN— LATITUDE OBSERVATIONS. 167 



revolution. The probable error of latitude from one pair of stars 

 — the part due to observation alone — for the two years was zti o".i2i 

 and =b o".ii2 respectively. 



The following work is an attempt to find whether the abrupt 

 deviation at irregular intervals of the observed latitude from the 

 mean value by an amount many times its probable error at Bethle- 

 hem is duplicated in a greater or less degree at Philadelphia. 



Thus, at Bethlehem the observations gave for the ten pairs 



The first values give a latitude agreeing closely with the finally 

 deduced value for that date ; therefore the difference, o" .24. repre- 

 sents a departure from the mean of 7 times the probable error. 



By applying the average difference between the tw^o stations 

 to the latitude at Philadelphia we can construct a diagram which 

 very nearly represents the efforts of two independent observers to 

 follow the motion of the pole at a given meridian. (Page 170.) 



The observations at each place for any day on which no corre- 

 sponding observations at the other place were made have been 

 omitted so that the diagrams represent about half the observations 

 at Philadelphia and two-thirds those at Bethlehem. 



The two curves agree pretty closely as to the lengths of the 

 periods and total variation, but the maximum and minimum values 

 at Philadelphia fall about three weeks earlier than the corresponding 

 phases at Bethlehem. 



Two methods will be employed in the comparisons. 



First Method. 

 The observations were divided into periods of about forty days 

 each and the means of the latitude for each of these periods found. 

 The simple difference was taken between the latitude for each group 



