157 



The surfaces S, might have been classified according as d = C. Tlie 



advantages of such a classification are not apparent in the analytical work 

 and can only be seen from the geometry of the surface or tlie generating 

 curve. In the vfork as presented the pseudosphere conies in as a special 

 case of the surfaces S i when d<2G, while if the classification had been 

 made as above indicated the pseudosphere d ^ O would be the dividing 

 surface in the classification. On the whole I think the classification 

 adopted is to be preferred. See figures V & VI for the different types of 

 generating curves d>C, d=C and d<C. The cut for d O is not given, 

 but a general idea of the curve may be obtained by leaving off the loop in 

 figure V. 



FfG.- J 



Fiq-r ^ 



