The following text may have been generated by Optical Character Recognition, with varying degrees of accuracy. Reader beware!

14 ted without this correction, 1 per cent should be added to any ordinate of that curve to get the correct reading. Steam for all the auxiliaries ex- cepting the blowers came through the starboard pipe only, in which a thin plate having a 2%4-in. orifice was inserted. The valve on the port line was tightly closed and the by-pass around the reducing valve on the aux- PUBLIBHEO BY AD. MACLACHLAN, 802 HOYLOTON BTREET, BOSTON. iliary line opened so that all the reduction in pressure was due to the orifice. The reduced pressures were read on the auxiliary steam gage on the engine room gage board and the flow of steam computed from co-efficients determined by experi- ments made at our engineering labora- tories. An independent orifice was necessary for the blower engines as this steam did not come from the aux- iliary line. The curve of auxiliary steam consumption shows the total from. these two orifices. The probable error in determining the. auxiliary steam is small. The so that THE Marine REVIEW | gage pressures were read to the near- est pound every 10 minutes and the average must be very nearly right. An error of 1 pound in this average would mean about 2 per cent error in the total auxiliary steam, but even if the error here were four times as great it would be of small moment, for the auxiliary steam is only about 1314 per cent of the total. The coal was determined by count- ing the buckets dumped on the floor. The points plotted on Plate 1 show the numbers of buckets at 10-min- ute intervals. Every hour the amount of unburned fuel on the floor was estimated, the corresponding points plotted, and the coal consumption curve drawn through them. The average net weight of coal was 506 Ibs. per bucket. Such a method as this for coal determination is open to question, but is perhaps the best that can be adopted on a short sea run. Even had the coal been weighed, the different condition of the fires at start and finish might have been con- siderable in so short a run. There is every reason to believe that any de- termination for so short a run would be somewhat too small. Our figures seem to show 80 tons for 12 hours equal to 100 tons for 15 hours, the running time from dock to dock, and this figure checks remarkably well with the amount of coal actually placed 'om board, an average being tak- -en over a large number of trips. ~~ NA NORRLA PETERS CO, WASNINGTONS ® F The quality of steam was de- termined by the throttling calorimeter, the sample being taken from a tee on the top of the steam main, inserted for this purpose where the pipe was tapped for the gage connection. Plate 2 shows the plots of engine test data. For the sake of simplicity only those observations appear in the plot which vary from the curve by an amount greater than the possible accuracy of observation. The close similarity of the curves and the small number of points through which the curves do not pass, is a strong check on the accuracy of the observations.