stylobate were perfectly horizontal, it would appear like the 
bed of a channel. The method of constructing the scamilli, 
and the application of them, is explained by a figure at the 
end of the book. 
Having proceeded thus far, the next object is to fix the 
bases of the columns in the places assigned to them. Their 
hollowed like the bed of a channel. Hence if we wish it to appear perfectly 
horizontal, we must make it to rise in the centre and fall gradually towards the 
extremities. Whether however it be that the line of the stylobate is interrupted. 
by being in some measure identified with the columns, and thus is only seen at 
intervals; or whether the eye becomes insensible, through habit, of the apparent 
curvature of objects which are presumed to be in reality straight, this great 
refinement, suggested by physical knowledge, does not appear to have entered into 
the execution of the works of the ancients. 
In rearing the columns upon a stylobate thus constructed, we must have recourse 
to some expedient for placing them so that their axes may be perpendicular to the 
horizon. This object may be effected by the use of scamilli impares, or smal 
plinths formed like wedges; placing them with the lower end towards the centre of 
the building. Scamilli of this description are adopted in the portico of Saint 
Martin's church; the stylobate of which is inclined towards the front for the purpose 
of carrying off the rain which occasionally falls upon it. These scamilli have their 
lower ends towards the pronaos, whereas those of Vitruvius have them towards the 
centre of the stylobate. 
This explanation of what was meant by the addition to the stylobate, will apply 
to the passage alluding to the mode of placing the capitals of the columns upon the 
shats. For the same reason that the line of the stylobate would appear concave. 
being below the eye, the line of the upper members of the episty lium would appear 
convex, being above it. To counteract this appearance it seems necessary to 
make the line of the episty lium concave towards the middle point; and the measure 
of the deviation from a straight line should bear a certain proportion to the adition 
made ia the centre of the stylobate. The same inconvenience would arise, with 
regard to the capitals, as before offered itself in fixing the bases, which might be 
obviated by the use of similar scamilli. Ifthe scamillihiere used be all equal in sige. 
it follows, that the abaci of the capitals will ise one above the other as they recede 
rom the central point of the epistylium: and consequently the capitals will be in 
different horizontal planes.