1897.] DOOLITTLE — VARIATIOX OF TERRESTRIAL LATITUDE. 435 



this occasion. In this connection I shall speak of the work which 

 we are doing at our newly erected Observatory of the University of 

 Pennsylvania. As the audience is not composed exclusively of 

 professional astronomers, it will perhaps be as well to give a brief 

 statement in explanation of the problem. 



The idea of possible variations in the latitude of points on the 

 earth's surface is by no means a new one. It is, in fact, as old as 

 Ptolemy. At various times, from that day to this, apparent changes 

 of this character have renewed the interest of scientific men in this 

 subject. It is probably superfluous to say that these supposed 

 changes of latitude were almost exclusively due to imperfect 

 methods of observation and to want of knowledge of various theo- 

 retical matters now well understood, that with improvements in 

 instruments, with more perfect knowledge of the effect of refrac- 

 tion, with the discovery of aberration and nutation these supposed 

 changes for the most part disappeared. 



More than a hundred years ago, the illustrious Euler gave to this 

 subject something like a scientific basis. In the development of 

 the law of rotation of a rigid body, for which we are indebted to 

 him, it was shown that a body like the earth, supposed to be per- 

 fectly rigid and in form an ellipsoid of revolution, unless originally 

 started in its diurnal rotation about an axis exactly coinciding with 

 the axis of figure, would have in addition to this diurnal rotation 

 another motion. Suppose the original axis of rotation to make a 

 small angle with the axis of figure. It was shown that this axis of 

 rotation would itself revolve about the axis of figure, describing 

 the surface of a cone, the angle between the two axes remaining 

 unchanged. The period of this rotation depends upon the princi- 

 pal moments of inertia of the earth or upon their relation to each 

 other, which may be found from the observed values of the con- 

 stants of precession and nutation. The resulting period proves to 

 be about 305 days, or ten months. 



If, therefore, this period has a real existence, that is, if the axes 

 of rotation and figure do not exactly coincide, it will, according to 

 this theory, be shown by a periodic increase and diminution of all 

 terrestrial latitudes in a corresponding period. Such a change was 

 not found, though sought for by many investigators. In fact, the 

 apparent perfection of the theory probably retarded the true devel- 

 opment of the matter for several years. However, about ten or 

 fifteen years ago, a number of apparent changes of latitude, found 



