PRESIDENTIAL ADDRESS—PRIESTLEY. 9 
In this Philosophy propositions are to be deduced from 
phenomena and generalized by induction. 
It is sufficient for us that Gravity exists and, acting 
according to the laws we have put forward, is adequate 
to explain the motions of the heavenly bodies and the sea.’’ 
I have given these extracts from the Principia to 
illustrate the change of opimion as to the function of 
Science. The change of method is brought out more 
clearly by a consideration of the history of the origin of 
Newton’s great work. Tycho Brahe had accumulated by 
observation a mass of data on the positions of the planets; 
Kepler, starting from the hypothesis that the path of Mars 
is an .epicyclic, failed to account for observed facts and 
formulated the elliptic hypothesis. This covered the facts 
as far as Mars was concerned, so he extended it by a 
tentative assumption that all the planetary orbits are 
ellipses with the sun at one focus. This generalization 
was found to fit the facts and was enunciated in 1609 as 
Kepler’s First Law. The second law appeared in the same 
year and the third ten years later. In 1638 Galileo pub- 
lished his Dialogues on Two New Sciences, which contained 
the results of his work on falling bodies. The immediate 
result was the concentration of the scientific world on the 
problem of Gravitation. 
Halley, Wren, Huygens, and Hooke all attacked the 
subject. Assuming tentatively that the planets were kept 
in their elliptic orbits by a foree of the same nature as that 
causing bodies to fall to the earth, and, simplifying their 
problem by assuming cireular in place of elliptic orbits, 
they deduced from Kepler’s Third Law that the attraction 
of the sun or earth on an external body must vary inversely 
as the square of the distance between the attracted and 
attracting masses. In passing we might notice here two 
characteristics of the new method; the attempt to find a 
single explanation of apparently different but possibly 
related phenomena; and the simplification of the problem 
by ignoring temporarily certain of the data, in this par- 
ticular case the ellipticity of the orbit. The first is in accord- 
ance with the accepted ‘‘Rule,’’ which was afterwards 
enunciated in the Principia, ‘‘No more natural causes are 
to be admitted than are sufficient to account for the phe- 
nomena’’; the second follows the practice adopted by 
