102 PHILOSOPHICAL SOCIETY OF WASHINGTON. 
vectores and the eccentricity of the ellipse, and directs us to sub- 
stitute the value of the perihelion distance. Euler does not, there- 
fore, give an explicit expression for the time in terms of the radii 
vectores and the chord, although he points out how this may be 
done. Considering the results that he obtained, Mr. HAut thinks 
that this theorem, with respect to all the conic sections, should be 
known by the name of Euler. 
Remarks on this communication’ were made by Messrs. Hr.1, 
Strong, and WINLOCK. 
33D MEETING NOVEMBER 2, 1887. 
The Chairman presided. 
Present, fifteen members and one guest. 
Mr. E. B. Etuiorr gave a brief description of a new form of com- 
puting machine, which prints and arranges in the usual form for 
addition any series of numbers and then prints their sum, the work 
of the operator being merely mechanical as in the use of the type- 
writer. Although designed especially for performing and printing 
work in addition, the machine may also be used for multiplication 
and division. 
Mr. Wm. HArKNEss presented a paper on 
THE CONSTANT P IN OBSERVATIONS OF TERRESTRIAL MAGNETISM. 
It was explained that this paper arose from a correspondence in 
Nature concerning the modes of computing the constant. In a letter 
published in the number of that journal for August 18th, 1887, Mr. 
Harkness referred to an expression for P given in Stewart and 
Gee’s Practical Physics, and suggested a more convenient form for 
logarithmic computation. In the number for September 8th, Mr. 
William Ellis gave another but somewhat less accurate expression 
for the computation of P; and in the number for September 29th, 
Prof. Arthur W. Riicker gave what he considered a more accurate 
expression than either of the others. 
Starting from the fundamental equations of Gauss, Mr. Harx- 
NEss derived a formula for P correct to terms of the second order 
inclusive. He then showed that Riicker’s formula, which pur- 
