NOTES ON RIVETS AND SPACING OF RIVETS FOR OIL-TIGHT WORK. 69 
assuming that the deflection of the plate edge caused by caulking probably varies 
as the ratio of the fourth power of the distance between rivets to the fourth power 
of the thickness of the plating, and that the extension of the rivet varies as the 
ratio of the distance between rivets to the area of the rivet. In order to make 
practical use of the above it was suggested that water-tight spacing be used for 
rivets in oil-tight work on naval vessels. 
A naval constructor stationed at the same yard carried the investigation 
further by considering the deflection that is caused by caulking and by the amount 
of general deflection of plating due to water pressure, and suggested that based on 
the assumptions made the following spacing would be satisfactory:—15-pound 
plate, 4 diameters; 20-pound plate, 4.5 diameters; 25-pound plate, 5 diameters; 
30-pound plate, 5.3 diameters; 35-pound plate, 5.6 diameters. 
These conclusions were deduced without reference to strength of joint, and it 
was pointed out that this consideration might be more important than questions of 
tightness. For five-eighths-inch rivets in 10-pound plating it was shown that approxi- 
mately .006 inch is the deflection or amount of seam that can be caulked off with- 
out exceeding the elastic limit of the material. The stress on the rivet for such 
caulking was found to be very moderate. 
The analysis used in deriving the above tables of rivet spacing for oil-tight 
work was made to confirm a theory that every unnecessary rivet is a possible 
source of leakage, and that for heavy plating a wider spacing would be permissible 
but considered only the stresses on the plating and rivets due to caulking to the 
exclusion of the more complex stresses experienced under service conditions. 
_ For thin plating, where large rivets in relation to thickness are used, the 
diameter is based more nearly upon theoretical considerations, but if the same 
considerations were followed in determining the diameters for thicker plates, the 
rivets would exceed the limit practicable to drive, and spacing based even on using 
the same number of diameters would at some point become greater than indicated 
in the tables. Thick plates, however, are used where greater strength is required, 
and as the rivets used are much smaller in relation to thickness a greater number 
is necessary to develop the strength of the plating. The uniform spacing in 
diameters adopted in practice, substantially, for all thicknesses of plating is a great 
convenience, and it is only necessary to increase the number of rows of rivets 
where greater efficiency of joint is desired. If the spacing suggested in the tables 
given is suitable for oil-tight work, it might be necessary to increase the number 
of rows to insure proper strength and friction of the joint, in which event there 
would be an increase in weight and the attempt to reduce the number of rivets 
defeated. 
