the revolutions per second corresponding to those times of one hundred revolutions and write them in the third column. Then from the plotted curve scale off the velocities in feet per second corresponding to those revolutions per second and write them in the second column. In using the meter, measure the time of one hundred revolutions and look in the table for the corresponding velocity. Meter rating tables are given in W. & L. E. Gurley's Manual of Engineers' Instruments.

The experience of the writer indicates that all the ratings of a meter when plotted on co-ordinate paper should give parallel curves. The distance these curves lie apart depends entirely upon the friction of the bearings of the meter wheel at the time the ratings were made. When a meter has been rated so that the curve has been accurately determined, all subsequent ratings should give the same curve or parallel ones, unless the meter has been altered in shape. All the different meters of one size and kind will usually give parallel rating curves. If the friction of the bearings only had changed the curves will lie apart from each other, but they will remain parallel. A slight change in the shape of a cup of the meter heel makes only a very slight change in the rating curve. Slight changes in the friction of the bearings make very slight changes in the intercept.

When computing the discharge, if the vertical velocity curve for that discharge section, has not been accurately determined the writer takes 96 per cent of the observed mid-depth velocity for the mean velocity for an open river. And for a river frozen over he takes 91 per cent. Each observed velocity should be multiplied by the cross section area extending from the point of observation half way to the next station on each side of it. To take the mean of two velocities and multiply the mean by the area between them is not as near correct as a low velocity may be given weight over too large an area, or a high velocity weight over too small an area, as would be the case when at one station the water was shallow and moving slow, and at the next station, one hundred feet away, it was deep and moving swift.

In order to do the best work with the current meter an engineer needs to have some experience, and needs to acquire some skill in the operation and care of the instruments used. He should be able to make all adjustments and repairs ever likely to be needed, or at least should know how to do so in order to know whether it is done correctly or not. When the writer has been stationed a long distance from a town he has had to stop in the middle of his observation and repair the stop watch, using a jackknife and a brass pin for tools; and by putting in a piece of pine cut from the skiff, and a piece of rubber cut from his boot for a spring, the watch was set going and the work went on again.

DISCUSSION.

Mr. T. T. Johnston: Mr. Price's paper is of a good deal of interest. It has covered the process of gauging streams by the use of current meter, and to some extent with the rod float, but there are one or two thoughts that occur to my mind in connection with the subject of the paper presented by Mr. Price. One was his remark with regard to the changing in the rating of the meter. Within the last few months I have had occasion to observe a meter of a well known type that was adapted for measuring the flow of artesian wells. The meter was rated about five years ago, early in 1893, and was rated very carefully, and since that time has been used nearly every month, once each month in about sixty artesian wells. The general use, covering a period of not more than one day in each month for five years, would make sixty days' use of the meter. It might be estimated that its use was not more than would happen in half a year's use in gauging a stream.

At low velocities it had not changed very much; at velocities which would cause eight or ten revolutions per minute, at velocities which would occasion revolutions of 300 to 400 per minute, the rate had changed 30 or 40 per cent. That is indicative to some degree of the sensitiveness of the instrument to change in the rating.

Another thought occurred to me in connection with the propriety of using a current meter in gauging certain streams in preference to using rod floats. Take, for instance, the work done by the sanitary district in Chicago, where the work done came as the floods came and very irregularly. In one year we had no occasion to make any measurements at all, but when they were made, they had to be made in great numbers and with great rapidity, and no preparation could be made for them. Under circumstances of that kind the current meter hardly seemed to be a desirable instrument, because probably it may not be kept in condition, or may not be ready for immediate use; a bit of sand may be buried in it, or something like that; a thousand and one incidents by which the instrument is affected may exist. It seems to me that the rod float is far preferable to using the current meter in such a case, and so when occasional measurements have to be made for flow of streams where a meter might have to be borrowed and hastily put in place, it would be far preferable to use rod floats, in my opinion. That brings up the question of the relative merits of rod floats and current meters in measuring the velocity of currents. The rod floats, if used properly, necessarily embrace a large proportion of the depth of the stream. Corrections have been suggested for the reason that the floats do not cover the whole depth of the stream. Figures of correction do not amount to much in relation to any use to which the information derived can be put, and are perhaps an unnecessary refinement. The rod floats embrace the velocity of the current

from top to bottom. Mr. Price's conclusion that the floats might move properly from one direction to the other on account of the boiling of the stream, it seems to me, applies as much to the current meter as to the rod float. It is a question in my mind whether after all the rod float is not more satisfactory in measuring the mean velocity at any vertical element in the cross section of a stream.

President Noble: I would ask Mr. Johnston whether, in the second reading of the meter, the readings were at all points slower than at the original points.

Mr. Johnston: No, at the low velocities the rating did not seem to change to amount to anything that we could check up. It was at the higher velocity that its rating changed.

President Noble: And it was slower?

Mr. Johnston: Yes, the meter showed some signs of wear, and that seemed to be the effect of the wear. It was clean and in good condition otherwise. It simply shows that the current meter is quite a delicate instrument, and its ratings must be very carefully watched.

XXXVI

ON THE USE OF COKE BREEZE IN SEWAGE PURIFICATION.

By JOHN W. ALVORD, Mem. W. S. E.

Read May 18, 1898.

The science of the purification of sewage has been of late years co-existent with the development of specialized bacteriology. Nevertheless important advances were made before the biological feature of the chemical reaction was fully understood.

Thus intermittent filtration was developed in England in 1870 by Dr. Frankland for the Rivers Polution Commission, essentially as it is used and understood today, but it was not until 1877 that the researches of two French chemists. M. M Schloesing and Muntz, established the fact that nitrification in sewage and in soils is the result of the action of an organized ferment.

Warrington, in England, presented the results of his investigations on this subject to the British Society of Arts in 1882, but it was very slowly that the researches of Pasteur and others forced the conviction upon sanitarians that the biological initiative of nitrification was a primary requisite and not a resultant feature.

Following this conviction the experiments of the Massachusetts State Board of Health conclusively confirmed the presence of the nitrifying organism, and developed much valuable information as to its proper environment. These experiments were commenced in 1888 and have continued to the present time, their more valuable features being published in the report of 1890.

It is from this point in the history of sanitary science that I desire to review more in detail the development of what, for want of a better name, may be called "Coarse Grain Filters," and notice briefly the most recent ideas advanced and experiments made in regard to them.

The experiments of the Massachusetts State Board of Health did not leave the question of intermittent filtration upon land in a wholly satisfactory condition from a practical point of view, although the facts they demonstrated enabled the sanitarian to prescribe with far more exactness the rates of flow and kinds of soil for different quantities and qualities of sewage than he had ever been able to do before. The most favorable results, however, were obtained through certain sands, not always found where needed, and necessary in such large quantities and areas as to render intermittent filtration unaccessible as a method of sewage disposal for the cities of the larger class.

Thus, Chicago, investigating the question about this time, was obliged to estimate upon sand beds in Indiana, the extent and

the Crossness outfall of the London main drainage. The ground was leveled, underdrained and filled with three feet of coke breeze. The sewage was flowed on to the bed to the level of the surface as quickly as possible, allowed to remain standing full at least an hour, and then drawn off with the least delay. Working in this way the filter passed 1 1-6 million gallons (imperial) daily for six days, resting one day. This is equivalent to about 1,200,000 U. S. gallons daily, including time of rest. The purification obtained was 78 per cent based on the oxidizable matter, and the filtrates were said to be clean and sweet.

In Sutton, England, a similar series of preliminary experiments were undertaken, on a more elaborate scale, but with approximately similar results.

Upon them as a basis a coarse grain filter was constructed by filling one of the precipitation tanks, previously used for chemical process, with burned clay. The sewage passed from thence into a second tank filled with coke breeze. The operation of this plant commenced Feb. 11, 1897, and a rate of flow of 773,000 gallons per acre per day was obtained, rest period included. The results of 76 days operation in the first tank gave a reduction of 66 per cent in the oxidizable matter, the effluent from the entire plant showing a total reduction in the oxidizable matter of 86 5-10 per cent.

The solid matters held in suspension were reduced by the first tank 95 per cent and by the combined operation 99 6-10 per cent. The total quantity of suspended solid matters in the sewage, which were disposed of by the bacteria, was equal to 77 tons of sludge removed during the time of treatment, the average sewage having contained 54 5-10 grains per gallon. The fact that this quantity disappeared, without cost or nuisance is striking evidence of the capability of the process. At times these filters were worked up to a rate of nearly 3,000,000 gallons of sewage per acre per day. Such high rates checked the bacterial action, but did not permanently disable the filters

At Exeter, England, coke breeze filters 5 ft. in depth, on a similar plan have been in use since Aug. 1896, receiving sewage effluent much more concentrated than average American sewage. The rate of flow is about 666,000 gallons (imperial) per acre per day, inclusive of rest periods, and the purification obtained 75 per cent in oxidizable matter removed.

In all of these biological filters the operation is essentially different from ordinary sand filters of intermittent filtration as now practiced. Indeed, these biological filters are not filters at all in the sense in which the word is ordinarily used. Prof. Dibdin calls them "Bacteria tanks." Another sanitarian proposes the name "Digestors." "Respirators" and "Aspirators' have been proposed." Biological Filters" is not perhaps inappropriate, although the word filter is not apt. In these biological filters, so called then, the tank is filled with sewage, preferably from the top. It