as a strong solution suddenly put in the boiler will not act to its full capacity or perform its work properly.

Lime. The lime, or Clark's, process is applied to, and effective in, waters containing much, or mostly, bicarbonate of lime and magnesia, and little or no sulphate of lime. Here the lime, in solution as lime water, unites with the carbonic acid of the bicarbonates and precipitates them, as well as itself, as insoluble carbonates. This precipitate with lime is double the quantity what it would be if caustic soda were used; but, in the absence of sulphate of lime, has the advantage of not leaving any carbonate of soda in solution. The drawbacks of the latter were alluded to above. Lime is detrimental in the presence of sulphate of magnesia, as it precipitates hydrated oxide of magnesia and leaves sulphate of lime, which is more harmful, in its place. Lime and Soda Combined. This is used when water contains both sulphate of lime and bicarbonate of lime and magnesia in such proportion that the amount of carbonate of soda necessary does not have its full softening effect. The treatment in this process is accomplished by the addition of the amount of soda necessary to precipitate the sulphate of lime, and the amount of lime necessary to react on the carbonates. The combined method gives results where either alone would do so only imperfectly. Water should be treated by this process only while Icold and in tanks.

Besides the above, there are other substances acted upon; but these are generally present in too small quantities to be considered. They are carbonate of iron, which is affected similarly to the carbonates of lime and magnesia and chloride of calcium, which acts like chloride of magnesium.

Barium salts are used to some extent. Caustic baryta acts like lime, and has no advantage over it. Chloride of barium precipitates insoluble sulphate of barium from sulphate of lime, leaving chloride of calcium in solution. This might answer where the water contains exclusively sulphate of lime, but it also forms a precipitate with sulphate of soda and sulphate of magnesia, unnecessarily increasing the precipitate and leaving the harmful chloride of magnesia in solution. The price of barium salts is, however, too high in comparison with other substances of equal or greater effectiveness to make their use general.

Sodium fluoride is practically the same in its action as carbonate of soda, but is much more costly. It precipitates lime and magnesia salts in a light, flocculent, non-adhesive consistency, and when pure has the advantage over soda of not leaving the water so strongly alkaline. As a good many commercial samples react strongly acid there should be some care exercised in its use.

Tri-sodium phosphate, or T. S. P., as it is sometimes designated by engineers. The value of this chemical consists in its ability to convert the soluble lime and magnesia salts in the water into, and precipitating them as, insoluble phosphates of lime and magnesia. These phosphates are of a very flocculent nature, having a specific gravity little above that of water, and on that account do not settle easily, but are continuously in suspension, and settle when the boiler is at rest as a soft mud that does not harden. Tri-sodium phosphate, on account of its alkaline nature, neutralizes any acids present which would otherwise cause corrosion, and an excess is not so likely to cause foaming or priming, as would one of carbonate of soda.

Bi-chromate of Soda.-The use of this chemical is patented in Germany, and has recently been introduced here. It precipitates scale-forming lime and magnesia salts as insoluble, non-scaling chromates. It is claimed that an excess of the chemical, even free chromic acid, has no corrosive action on the iron or packing.

Tannin or tannic acid is used to some extent. It forms nonscaling tannates of lime and magnesia, but attacks the iron, and is not to be recommended.

Tannate of soda, like tannic acid, forms tannates of lime and magnesia, but is much safer to use.

Sugar is also sometimes used, precipitating the saccharates of lime and magnesia.

MECHANICAL COMPOUNDS.

A good many chemically inert substances are used as water purifiers, but they accomplish their object only partly, if at all. Such are:

Sawdust. This is supposed to furnish a nucleus or center for the crystallization of the hardening salts preventing their uniting together.

Mucilaginous substances, starch, potatoes, etc., have the opposite effect to that of sawdust, that is, to envelop or surround the minute crystals of scale, and thus prevent them from hardening. Such substances, however, cause foaming and priming.

TO PREVENT CORROSION.

For the prevention of corrosion an iron-zinc couple is frequently used and seems to give good results. This is obtained by attaching zinc plates or rods to the iron bracings of the boiler, whereby a galvanic action is set up. The corroding substance attacks the zinc, which can be easily and cheaply replaced, leaving the iron of the boiler practically intact.

GENERAL RULES.

In the proper selection of boiler compounds or water purifiers the following considerations should be observed: Never use any of the so-called universal compounds, alleged to be good for any scale or any water. There are numberless small concerns or individuals going into this sort of "business" every year who claim to have the best compound ever produced and who recklessly condemn that of every other competitor. All they ask for is a trial of their compound, which means neither more nor less than experimenting with your boiler at your expense and risk. The greater part, in fact nearly all, of the secret compounds now sold are nothing but a mixture in varying proportions of some of the chemicals described above, usually sold at from three to twenty-five times their actual cost, the basis of most of them being soda, colored or blended in every imaginable manner.

Deal only with reliable concerns of financial standing, such as will actually make an analysis of your water, and furnish you a duplicate of the analysis, and who will prepare a compound to suit your particular water.

TRANSMISSION OF POWER.

In order to explain what is necessary to determine when shafting has to be erected for the transmission of power, we will take the most commor. case.

An engine of twenty-five horse-power is to be connected to a shafting supplying, by means of three pulleys B, C and D, power to different machines (Fig. 54). The machine connected to B requires five horse-power, the machine connected to C requires eight horse-power, and the machine connected to requires twelve horse-power. Diameter of pulley B = 6"; diameter of pulley C = 18"; diameter of pulley D = 20"; revolutions of engine = 75; revolutions of shaft = 100. The diameter of the pulley A' on the engine is = D and 32′′. The diameter of pulley 4, width of all belts, and the diameter of shaft are to be determined; also the bearings to be placed at the proper places.

DIAMETER OF PULLEY ON ENGINE,

Calling the respective number of horse power to be transmitted h, h', h", . . . and the radii of the corresponding pulleys r, r', r", . . . we have

The diameter D of pulley 4, which is to be driven by pulley A', calling the revolutions of the engine

N, and of the shaft = n,

The width of the belt required for pulley A we find, as explained before, by ascertaining the frictional surface of the same for the belt. Referring to Figs. 54 and 55 we find this surface to

be c' ď. We must, of course, examine the smaller pulley, as this has the smaller frictional surface. If we do not want to find this surface by drawing the plan, we can proceed as follows: The total circumference of pulley A is dπ

the arc of 360°, and to find the number of degrees for c' d', draw two vertical lines through the center of the shaft of each pulley A and A to the line connecting both centers. We know that angles oec =ofd = o' e' c'′ = o' f' d' = =w, and drawing a line through e'

parallel to oo' we know that angle g e' e is also was the sides of both are vertical to each other. Therefore

and as per table of natural trigonometrical functions w = 2°, or, in this case, a very small one which we could have neglected and taken simply one-half of the periphery of pulley A. However, the angle wanted is 180 4 and the arc 180-4 37 d π 37", and expressed in parts of the radius =—.

360

The formula for the friction was P

12

Qemw, and to get P we have to insert the values m and w, and then the table of sizes of belts and tensions will give the required width of belt. m = 0.12 for old greasy belts.