198 R O L L I N 
wedge of one inch thick to be introduced into every joint 
after the rings are screwed together by the joint-plates of 
wrought iron, which unite the fellies. These plates are to 
lay upon the plain surface of the felly, and not to be let in as 
the common wooden rings of water-wheels, in order that the 
oak wedges may completely fill the joints at the ends of the 
fellies. The wheels are to have wooden arms, and it must 
be observed, that the mortises through each of the iron 
fellies, for receiving the ends of the arms of the wheels, are 
to be about two inches and a half in width, and that they 
are to be a little dove-tailed, in their length only, so that 
the mortises being longer on the outside of the ring, and 
the wood of the arm being spread into them with wedges, 
will produce firm ties to the centre; but as a farther security, 
pins are to be put in after the wedging is completed. The 
mortises in the rings for the starts, which support the float- 
boards, are to be four inches by two, without dove-tailing, 
or rather they should be larger outside than inside. The 
breast or float-boards will fix by nails into the joints and 
arms, where they fall; but that the breast-boards for the 
intermediate floats may also have a fastening, holes of about 
one inch diameter, and about four inches deep, must be cast 
in the ring, at the places for every other float; these holes, 
being filled up with pieces of oak, will afford places to drive 
the nails for securing the boards. The axles of the water¬ 
wheel are to be of cast-iron, with flaunches to screw the arms 
of the wheel against. The total length of each axle is to be 
seven feet one inch, and the diameter throughout is to be a 
circle of nine inches. 
The brasses upon which the necks of the water-wheel axles 
are supported, are intended to be let into cast-iron stocks, 
which are again supported upon wooden bed-planks, and 
those upon the cap-stones of the walls, which (under these at 
least) are supposed to reach all across the thickness of the 
walls, those necks being first truly and smoothly turned; at 
each end, beyond the neck, is formed an astragal or 
moulding, to keep the wheel in its place from moving end¬ 
ways. The ends of the axles are terminated by an indented 
head, shaped somewhat like a square citadel in fortification, 
and an iron box is fitted upon this to communicate the 
motion to the rollers, the surface lines of the indented head 
being formed a little rounding, that the box may not only 
be certain of taking its bearing in the middle, but likewise 
be capable of complying with the motion of the rollers; 
and in order to give still more liberty, the end of the box 
which is farthest from the water-wheel, is formed into a 
square of eight inches, which is again surrounded by another 
box, whose external surface is round. This box is formed 
at the end of a round spindle, or axis, three feet long, and 
terminated at the other end with a square of eight inches, 
which enters one end of a square box, and at its other end 
receives the square of the roller, supposed to be of six inches, 
but may be of any other size which is thought necessary. 
It is to be noted, that all the squares are to be made larger 
than those of the rollers, in order that they may wear longer; 
and all the insertions are to be less than those at the end of 
the water-wheel axles, that the axis may not be rendered 
useless by the wear or failure of the citadel heads which are 
introduced, as they are expected to last many years; and if 
any thing should happen to them, the axles are made alike on 
both sides of the water-wheels, that they may change ends; 
therefore, there is nothing of consequence likely to fail by 
wear or breakage, except the smaller intermediate work 
between the axles and the rolls, which is easily replaced. 
Holes are to be made through the boxes and joint parts for 
iron bolts to pass through rather loosely, so as to prevent the 
boxes and squares separating, but not to confine the joints 
from yielding to the motion of the rolls. The water-wheels 
are supposed to be closely adapted to their conduits, and 
their axes to remain immoveable as to height, at the dif¬ 
ference of eleven inches in level, while the rolls are supposed 
to vary in their diameter from twelve inches to nine. This 
will be allowed for, by the distance that the squares upon 
the rolls are from the end of the axles ; for though the house 
is supposed no more than ten feet wide between the walls, 
G - M I 1 L. 
the distance between the middle of the citadel heads at tire 
end of the water-wheel axle, and the middle of the squares 
of the rolls respectively, is upwards of four feet; and in that 
length no less than four joints are introduced, every one of 
which complying a little, a small difference in the height of 
the roll will produce no sensible difference in the com¬ 
munication of the motion from that of a right line ; all the 
joints being kept oiled or greased, which will not be less 
proper on that account than to keep the parts from wearing. 
In order to preserve the directions as near as possible to a 
right line, Mr. Smeaton proposes that the lower roll shall be 
placed originally half an inch below the centre of the axis 
which turns it: suppose the rolls were first made of twelve 
inches diameter, while the difference of the height of the axis is 
only eleven, the upper roll will be just half an inch too high, 
so that the compliance in each will be equal, and no more 
than half an inch in four feet length. Thus it appears, that 
when the rolls are reduced by wearing from twelve inches to 
eleven each, then the upper roll as well as the lower will be 
half an inch too low for its axis;, still neither will need to 
comply or vacillate more than half an inch. The roll being 
now of a just diameter to answer the different heights of the 
axis, let the lower roll be raised to its just height, and then 
both the rolls will work true till they are further reduced; 
but when they become reduced to ten inches and an half 
diameter, the upper roll will become half an inch too low; 
then raise the under roll half an inch above the line, and 
the upper roll will then be truly in the line; so that when 
the upper roll is come down to be half an inch below the 
line, the rolls will be reduced to ten inches; then raising the 
under roll a quarter of an inch more, it will be three quarters 
above the line, and the upper roll will be a quarter of an 
inch under the line; and when it comes down to three 
quarters below the line, the rolls will be reduced to nine 
inches and a half; therefore, lastly, raise the under roll 
another quarter, so as to be an inch above the line, and the 
upper roll will be reduced to half an inch below it, so that 
when it is come down to an inch below it, the rolls will be 
reduced to nine inches. In this way the departure of the 
rolls from a right line will never become more than half an 
inch, while they are reducing from twelve to ten inches; 
nor more than three quarters, while they are reducing from 
ten to nine and a half inches; at the worst they will be no 
more than an inch, while they are reducing from nine and a 
half to nine inches. The greatest inequality is purposely 
made at this place, both because the purchase of the wheels 
is then greatest and most able to overcome an addition of 
friction, and because the time that they will continue in this 
state is the least. If the axes are placed at ten and a half inches 
distance instead of eleven, the vacillation each way will never 
exceed three quarters of an inch ; nor more than one inch 
to reduce the rolls to eight and a half diameter. 
The raising of the under roll is not to be done by raising 
the whole of the bed of the roller-frame; this is to be set 
originally half an inch lower than the true line ; and when 
the lower roll requires raising, it is to be effected by putting 
iron plates under the carriages of the lower roll gudgeons, 
so that they will stand as much higher than before, and not to 
make the several rises by additional plates, but to have plates 
of the different gages, so that each will lay in one solid piece. 
Mills, on this construction, are still used in many iron¬ 
works for rolling coarse iron bars, but are unfit for any better 
purpose, from the difficulty of adjusting the two water- 
wheels to an equal velocity; and if one roller moves quicker 
than the other, the metal becomes more extended on that 
side than upon the other, and is thus rendered convex. 
Another defect is, the want of proper fly-wheels to regulate 
the mill; for the cast-iron rims to the water-wheels by no 
means answer the purpose of fly-wheels, unless they are 
made to revolve so quickly that the water loses much of its 
effect upon the floats. Fly-wheels are, perhaps, more useful 
for rolling than in any other kind of mills, because the resist¬ 
ance to be overcome is so variable; being at one moment very 
great for a large piece of iron, then smaller whilst it is passed 
through a second time, because the iron is to be less reduced; 
