BOILERS FOR MARINE PURPOSES.

71

altogether independent of the factor of safety employed. I meanthe strains brought upon the boiler by the unequal expansion of itsdifferent parts. Ordinary wrought-iron plates, if left free fromstress, expand '0000064 of their linear dimensions for each degreeFahr, increase of temperature. Also if the plates are subjected tostress they alter in length a certain amount according to the qualityof the iron; the more ductile irons altering more for the sameamount of stress. Taking as the mean value of E, 29,000,000 (thevalue given by Rankine) we find that a stress of 186 Ibs. per squareinch will give the same alteration in length as i° Fahr. If, now,the ends of a plate are rigidly fixed so that it is incapable of alteringits length, an increase of i°Fahr. will subject it to a compressive stressof 186 lbs. per square inch, and a decrease of i° to a tensile stressof equal intensity; and it is to be observed that these stresses aretotally independent of the sectional area of the plate. Now, in thecase of a furnace, the portion above the fire, especially when coatedwith even the thin enamel or scale which is necessary to preserve itfrom corrosion, must be considerably hotter than the portion belowthe bars. Hence the top of the furnace tends to get longer thanthe bottom. If the end fastenings of the furnace were so rigid asto maintain the top and bottom of same length, the top would haveto be compressed and the bottom stretched, and every difference ofa degree Fahr, in the temperature would produce a compressivestress in top and a tensile stress in bottom of 93 lbs. per square inch.But actually the end fastenings are not so rigid, and the strainscaused by the unequal expansion are not distributed from top tobottom by the ends only, but also in a great measure by the resist-ance to shear of the plate, and hence the greatest stresses come atthe middle of the length of the furnace. Also, it is evident thatthese strains are not uniformly distributed, and hence their maximummust be greater than their mean, and with a great difference oftemperature the stresses reach a high figure. The only way tostrengthen furnaces from such strains is either to prevent the differ-ence of temperature, or eise to allow the crown freedom to expand.”

The question of reduction of strength of plates by punching ordrilling has had much attention, and experiments go to prove thegreater strength of drilled plates. Steel plates should always be .drilled; if punched, they must be annealed afterwards to reduce thelocal strains set up by the action of the punch.

The corrosion of boilers is an important matter, and recently since