carbonate of lime, more sulphate of lime and carbonate of magnesia. The superheating of the plates from the cause stated would possibly lead to decomposition of a film of water next to them and account for the corrosion, as oxidation under these circumstances would seem bound to occur ultimately.

It is not in place here to present chemical conundrums, and the author is not prepared to state whether the firebox and boiler flue corrosion was due to this cause or the carbonic acid in the well water. Before treatment the water badly corroded the supply pipes or water main leading to and from the supply tanks. I believe this corrosion was due to the carbonic acid in the water, as there are several instances which have come to notice at other points where corrosion in iron conduits of otherwise excellent water are producing disastrous results which can be traced to no other source as easy of explanation.

A switch engine with clean boiler was put to work on the wharf at Port Los Angeles, using only treated water, not long after the plant was put in operation, and observations made at intervals on the action of the water. No scale forms from its use, but a small deposit of magnesia sludge, having no tendency to form scale, has at times been observed. Otherwise the treatment has greatly improved the quality of the supply.

The author has done considerable experimenting in his laboratory, using different reagents and methods for the purpose of treating waters for steaming purposes. All samples were analyzed after treatment, and the double method described seems to have been much the more satisfactory in result.

As to the chemicals used for treatment, there are other reagents in the market for the purpose. Fluoride, phosphate, and hydrate of soda are among these and with which experiments were made, but the market price of these materials is not comparable with the dry carbonate, or soda ash, for the results produced. If, for instance, we wish to reduce the sulphate of lime in a well water (permanent hardness), it is practically immaterial whether the precipitate formed is fluoride or phosphate of lime, the neutral sulphate of soda (or potash) will remain as formed in the chemical exchange of bases. The phosphate acts somewhat quicker and the precipitate is heavier and settles sooner, but its cost is incomparable with soda ash (90 to 94 per cent. absolute Na CO) at less than two cents a pound.

As concerns the " temporary hardness" in natural waters

produced from bicarbonates of the scale forming bases, there seems to be no better or cheaper method of reduction than by use of quicklime. No reagent obtainable at the price will absorb carbonic acid more greedily. It is itself precipitated as carbonate of lime by the carbonic acid which it has taken up, and in thus depriving the naturally contained carbonates of their solvent allows them to precipitate also. Besides this, it will throw out of solution iron and silica to a certain extent as well as organic matter.

Some waters, from absence of incrustating sulphates, do not require the double treatment. For such stations the lime treatment alone is sufficient, and one tank, Fig. 97, with its adjacent circulating tank, Fig. 96, is made to serve.

The "standardizing" of the process for treatment of such waters as may thereby be benefited has been based on the following principles:

All tanks, vats, etc., to be made in sets of three sizes, the smallest being adapted to water stations having a delivery of 4,000 gallons per hour or less, the second 4,000 to 8,000 gallons, and the third 8,000 to 12,000 gallons.

Storage and settling capacity for a station to be sufficient for at least 24 hours' supply, unless circumstances allow other arrangement as to settling.

The character (from analysis) and rate of flow of water at a station being known, it remains to substitute such weight of reagents in a circular of direction, properly posted at the water station, as will supply necessary instruction to a person of ordinary intelligence.

DISCUSSION.

Mr. Albert A. Cary.-The purification of water for boiler use is a most important subject, which, I regret to say, has been sadly neglected by American engineers, and, in consequence, we find steam boilers all through this country suffering from pitting, corrosion, and scaling, which results in more or less rapid deterioration and increase in fuel bills, to say nothing of the worry and trouble connected with the attending repairs made necessary, and the laborious work required to keep the boiler in running condition.

This state of affairs is, in a majority of cases, unnecessary, as there are very few of the troublesome boiler waters which

cannot be treated in a comparatively simple manner by either mechanical or chemical means so that the boiler will be supplied with, practically, a pure water, and where water has been troublesome in boilers, such purifying equipments will be found to be one of the best paying investments in the entire steam plant.

Undoubtedly, the best method of treating boiler waters is to rob them of their scaling or corroding matters before allowing them to enter the boiler, but in some plants the equipment necessary for this purpose is regarded as too costly, and then there remains simply "the next best thing to do." Following this, and without proper advice, we find the boiler user often the victim of the many boiler quacks with both mechanical and chemical cures for all boiler troubles, who "treat" the water after it is fed into the boiler. All that the honest treatment of water inside of the boiler consists of is simply the use of some substance which will keep soft such scale as tends to form a hard, resisting crust, until it can be blown out, or removed by some continuous process, or else the use of some apparatus which will draw out continuously, or intermittently, the scaleforming matter before it has time to form a hard crust. These results can be obtained in the usual partially perfect way by the use of the cheapest chemicals or apparatus.

Many of the co-called boiler compounds are composed of these very cheap chemicals, but they are sold at prices ranging from five to twenty-five times their actual value, and as many of them contain positively harmful ingredients, I would advise boiler users to adopt them only after consultation with an experienced chemist or expert.

By describing this method of internal treatment of boiler feed waters I do not wish to be understood as endorsing the method, although I have been forced to adopt it in certain cases. I believe that a boiler has all it can do to furnish the steam required without converting it into a chemical laboratory, as ill results are pretty sure to follow such misusage sooner or later. External mechanical treatment of boiler waters depends principally upon the fact that when water contains certain substances in solution, and when the temperature of such water is raised above a certain degree, more or less definite, this dissolved matter cannot remain longer in solution and, therefore, it is precipitated and held in mechanical suspension. This

precipitate, in one treatment, is deposited into the top, or onto the bottom or sides of catch-pans, where it is supposed to remain, the clear water passing onward into the boiler.

Filters are also used to remove all kinds of mechanically suspended matter, whether existing originally in the water or else being precipitated by some process from a chemical solution.

In the paper before us we find an external treatment by chemical means, which is very effective when properly designed and manipulated.

Mr. Stillman speaks of basing his process upon the Clark or Porter-Clark process. One might infer from what he has said that both of these processes are identical, but this is not so. The Clark process was originally a simple lime process; the lime-water (not cream of lime) is mixed with the water for treatment by the simplest means, and the precipitated carbonates were allowed to settle slowly in large and expensive tanks. The sulphates of lime and magnesia present in the water were not eliminated by this process. Mr. Porter followed Mr. Clark's invention with an arrangement by which he mixed the limewater with the water for treatment, by use of revolving paddles, and then, without waiting for the precipitate to settle, he forced his water, with its mechanical solution, through a filter press, from which was delivered practically clear water.

In the method described by Mr. Stillman he treats his water in two separate processes, first mixing his cream of lime in the water for treatment, to throw down the carbonate of lime and carbonate of magnesia, and afterwards he mixes the carbonate of soda (or soda ash) with this partially treated water to throw down the sulphates as carbonates.

To those who are not familiar with these chemical reactions in the treatment of boiler feed waters, who care to follow this subject, I would refer them to a popular article which I wrote, and which appeared in the March, April, May, and June numbers of 1897 in the Engineering Magazine.

I have been quite successful in treating similar water to that described by Mr. Stillman, by a single-process method, using, instead of the cream of lime, lime-water, which I have found very much more satisfactory, and instead of using the carbonate of soda, I have used caustic soda, which acts quicker than the carbonate of soda, and seems more powerful in its effect. Probably (when acting upon the sulphate of lime) it more readily