[ 534 ] 
The thicknefs of the fpindle then being determined 
by thefe conditions ; it is alfo manifeft, that, to pre- 
vent the fpindle from being palled out of its place 
by the draft of the rope, there muft be a fhoulder 
formed upon it, which muft be greater than the part 
of the fpindle before defcribed, for the fpindle to bear 
againft. The fize, that Mr. Saumerez propofes to 
give to his Y, is 27 inches the whole length ; r y 
inches for the length of the arms (which are to be 
opened to a right angle) ; 8 inches for the length of 
each vane ; 4 inches and an half broad, and the items 
and fliank to be two-thirds of an inch thick. Ac- 
cording to thefe dimenfions, the reftftance, that this 
part of the machine will meet with, in pafting thro’ 
the water, will, in the fwift motions of the fhip, be 
very confiderable : confequently, the neceflary bulk 
of the pivot-end of the fpindle, and its fhoulder, will 
occalion a confiderable fridtion in the turning thereof, 
and retardation to the rotation of the machine. 
To cure thefe defedts, as much as poftible, inftead 
of the Y before defcribed, I made trial of a fingle 
plate of brafs, of about ) o inches long, 2 and an 
half broad, one-thirtieth of an inch thick, and cut 
into an oval fhape. This plate being fet a little atwift, 
and faftened by one end to a fmall cord, in the man- 
ner of the Y, is likewife capable of making a ro- 
tation, in being drawn through the water j but with 
this difference, that as this is but a fmall thin plate 
drawn edgeways through the water, its refiftance, in 
pafting through it, is much lefs ; of confequence, a 
much fmaller line is fufftcient to hold it, which again 
conftderably diminifhes the refiftance ; and this, of 
courfe, proves a double diminution of fridtion in the 
fpindle : 
