580 Proceedings of the Royal Society 
3d, These diameters are all equal, or if they are unequal, their 
inequalities sensibly compensate one another [in averages]. 
Constitution of Gravific Corpuscules. 
1st, Conformably to the second of the preceding suppositions, 
their diameters added to that of the bars is so small relatively to 
the mutual distance of parallel bars of one of the cages, that the 
weights of the celestial bodies do not differ sensibly from being 
in proportion to their masses. 
2d, They are isolated. So that their progressive movements are 
necessarily rectilinear. 
3d, They are so sparsely distributed, that is to say, their dia- 
meters are so small relatively to their mean mutual distances, that 
not more than one out of every hundred of them meets another 
corpuscule during several thousands of years. So that the unifor- 
mity of their movements is scarcely ever troubled sensibly. 
4:th, They move along several hundred thousand millions of 
different directions ; in counting for one same direction all those 
which are [within a definite very small angle of being] parallel to 
one straight line. The distribution of these straight lines is to be 
conceived by imagining as many points as one wishes to consider 
of different directions, scattered over a globe as uniformly as pos- 
sible, and therefore separated from one another by at least a second 
of angle; and then imagining a radius of the globe drawn to 
each of those points. 
5th, Parallel, then, to each of those directions, let a current or 
torrent of corpuscules move ; but, not to give the stream a greater 
breadth than is necessary, consider the transverse section of this 
current to have the same boundary as the orthogonal projection of 
the visible world on the plane of the section. 
5th, The different parts of one such current are sensibly equi- 
dense ; whether we compare, among one another, collateral portions 
of sensible transverse dimensions, or successive portions of such 
lengths that their times of passage across a given surface are 
sensible. And the same is to be said of the different currents com- 
pared with one another. 
7th, The mean velocities, defined in the same manner as I have 
just defined the densities, are also sensibly equal. 
