A recent article by Lieutenant A. F. H. Yates, U. S. Navy, on the operation and management of the Parsons turbines on the Chester* possesses particular interest by reason of its very careful presentation of those little details, gleaned from actual experience, which were found to be most important in the care and handling of this comparatively new type of machinery.
While realizing that the Parsons turbine is no longer a novelty and that the principles of its action, as well as the main feature of its construction, are widely known, it is nevertheless still quite too limited as actual installations in our navy, to have, as yet, afforded to any large percentage of the officers opportunity for Personal study or observation of the assembling of its various elements; and these notes, therefore, are submitted as being pertinent to a better general comprehension of the construction of the engine and the several arrangements (actual and proposed) for the most efficient distribution of the power to suit naval conditions.
The half-tone illustrations here used are of the low pressure and ahead turbine of the Preston, this happening to afford a more comprehensive picture of the different stages than would any other single turbine, and at the same time, by using photographic views, we are assured of securing every obvious detail without betraying any confidences in actual dimensions of parts or approaching an exhibit of structural plans.
*May, 1909, issue Journal of American Society of Naval Engineers.
It is quite impossible to conceive of a steam engine composed of more separate parts than is this turbine and the first thought of the engineer in taking up the question of construction, is one of concern regarding the security of the combination. In spite of the excellence of the methods adopted for insuring the integrity of the assembled units, some anxiety on this score follows the user, perhaps to a greater degree than in the case with the reciprocating engine. This is not alone due to the multiplicity of the individuals elements but also to concealment of the working parts and to a knowledge of the close range of clearances used and the unequal effects of high temperatures upon the different materials comprising the structure. These tend naturally to invoke a mental tension on the part of those responsible for the machinery and its performance unless they be of phlegmatic mould and calmly anticipate nothing. However, with that care in building and watchful skill in management that are easily possible to secure there should neither be apprehension of nor actual disablement in usual service. This is not a case of simply admitting steam and trusting in Providence, but rather recalls the statement of the man who, in describing how his family was saved from disaster while out driving said: "Yes sir, the horse ran away and the reins slipped from my hands and if it had not been for Providence and another gentleman, who stopped the horse, we'd all have been killed." We have to have the "other gentleman" ever on the qui vive.
As to the great number of pieces going to form the Parsons turbine we have in the blading alone some startling figures. For instance in the low pressure and astern turbine of the Preston (illustrated) there are over 63,000 blades in casing and rotor, each blade of course having its distance piece or calking piece behind it, and these, with the blades, making a total of more than 126,000 separate pieces fitted in the grooves of drum and cylinders of this one unit, and over half a million in the entire set of turbines for this boat. In the Utah's outfit there are over 825,000 blades alone, ranging from 7/8-inch to I3½ inches in exposed length and requiring about 52 miles of blading strips for their manufacture.
Regarding clearances and effects of high temperature we must remember that the casings are made of cast iron, the rotor drums of forged steel and the blades of brass. Also that, in order to reduce waste in leakage over tips of blades it is extremely necessary to bring the tip clearance of blades from casing or drum to a minimum and that this clearance allowed cold is further diminished by the greater expansion rate of the brass.
It will be obvious to the reader, after a brief study of the subject, that the greatest care must be exercised in warming up the engine lest we distort the apparatus and cause binding or bearing of blades to their injury or destruction. As yet we are not using superheated steam but the pressures are fairly high; that of the Preston being 255 pounds absolute at engine corresponding to a temperature of over 380°F.
Of course the casings are not steam jacketed because the contact “walls" in the engine are principally blades, but possibly a steam jacket for simply warming up the casings first, would be an advantage in obviating troubles from unequal or distorting expansion of rotor in reaching final temperatures. The casings are massive and proportionally slower and more difficult to heat safely and proportionally, with the rotor, by simply admitting steam within. The low pressure turbine casing of the Utah, in the rough, weighs about forty tons.
To keep in mind the values of clearances and thrust tendencies we must use the typical diagram of turbine blading preliminary to a discussion of the assembled machine. Fig. I gives a general idea of the relative proportions of tip and side clearances of blades and it is also important to note several other features illustrated here: the uniform section of the blades, the thicker edges of blades always being the receiving edges of steam flow, the binding wire and lacing, the side grooves in the main grooves of rotor and the obvious effect of steam thrust on the rotor blades to counteract propeller thrust.
The blading is made in strips 16 or 18 feet long—of brass (tin-copper) drawn or extruded through dies which correspond to the exact section required for each expansion, and these strips are then fed to a machine which cuts the blades to the exact length desired and at the same time stamps in the two calking grooves E and cuts the notch F for binding wire G. Until recently this notch was cut by a small machine saw after the first operations. The top edge or tip of the blades are also ground to a thin edge as indicated in sketch.
The distance pieces, packing pieces, or calking pieces, as they are called C, which space the blades in grooves of rotor and casing, vary in section according to the pitch, angle and section of the blades of each expansion, and this is taken from the design of blading arrangement, and conforms to the exact shape of the space between two blades, so as to fill, when calked in, the entire groove. The pieces are cut off to the depth of the groove and the blading proceeds as follows: For rotor one of the packing pieces is placed in the first and "calked in" by calking tool on its upper face, expanding it sufficiently to fix it against sliding in the groove when driving up the following members. Then several blades and packing pieces are inserted alternately behind this "stopper" and driven hard up against it by hammer and a special tool shaped to fit the back of the packing piece and curved to "feed" properly into the groove. The blades themselves are lightly tapped down to insure bottoming at the time before final driving and then another group inserted and driven home, this driving being sufficient to wedge the pieces in tightly enough to prevent any displacement by turning the rotor around. At the insertion of the final blade of a row it is usually necessary to file off lightly the back of a packing piece to provide for tapping in the last blade tightly, and then the final calking is done by use of a long calking tool fitted to slip down between the blades and cover the top face of the packing piece. This tool has a small "center" which leaves the imprint on the top of each piece calked and thus affords means of detecting quickly, by simple inspection, any piece "skipped." The calking is rapidly done and makes the pieces fill in both the small side grooves shown in the rotor groove and also the grooves on back of the blades, locking the whole mass effectively in place.
The rows of longer blades are now further stiffened by binding them with a continuous strip of brass wire, or rectangular section, fitted into the slots F and lashed in place by soft copper wire as shown. This lashing is secured by silver solder and blow pipe, not only to the binding wire but also to the edges of the blades where it rounds them. Sometimes in very long blades two binding wires are used equally spaced on length of blades while in the very short blades no binding strips are used.
In blading the casing very much the same method is followed, only that, in this case, the top guide blades in each row of each half casing are heavy and countersunk into the casing, forming the "stoppers" for driving up the blades and distance pieces.
In large turbines such as the Utah's, this method of blading is modified in order to expedite and facilitate the work. The manner in which this is accomplished is by first fitting up a cast iron "former" representing a portion of the grooving of each step. These formers are made in two parts bolted together, the division running through the middle of, and parallel with, the groove, so that at any time when the wear due to fitting many blades increases the width of the groove the former can be taken apart and refitted to the exact width.
In using this former the blading and distance pieces are made up in sections of about one-tenth of the the circumference of each step. Each blade and each distance piece is accurately drilled near the bottom to accommodate a bottom binding wire, varying from one-sixteenth to one-eighth-inch diameter. This wire is inserted in the first distance piece and soldered to it, and the following blades and distance pieces are then strung on the wire, and as they are strung they are driven solidly home against a stopper at the extreme end of the cast iron former, exactly as though the work was being done in the regular grooves of turbine. When the section is completed the other end of the bottom binding wire is soldered to the last blade and the regular binding strip at the outer end of the blades is inserted and secured as previously described; so that when the section is removed from the former it is firmly united and ready to be placed in its proper group. After this the only caulking necessary is to expand the distance pieces into the grooves of rotor or casing. This being readily accomplished by using an offset caulking tool. In this way the blading of a turbine can be completely set up before the rotors and casings are ready for them.
It should be observed that the outer binding strip of each section is left long enough to solder to the strips of the adjacent sections after being fitted in place on the rotors, so as to afford the additional security of a tire to the revolving blades.
The complete rotor of Preston's L. P. and astern turbine is shown in Figs. 3 and 4, and the lower half of casing for same in Fig. 2, the lettering in each indicating corresponding or mated parts, and the peculiarities of the blading may be best followed further in these views at this time.
In Figs. 3 and 4 the first six "steps" of blades (36 rows) from reader's right comprise the L. P. ahead turbine rotor, and the binding strips can be seen as a little beading on the outer edges of the blades. There are about 18,000 blades in these six steps on rotor alone, ranging from 1? inches to 8? inches in length and, while there are only six steps, there are nevertheless eight expansions, for in the last step, counting 12 rows, each successive four rows are of blades with different spacing and angle giving three separate expansions in this step. On the left end of the rotor are seen the astern blades A, which by reason of using high pressure steam are much shorter. Here we have only three steps but also four expansions, each of eight rows, and the blades run from 1 inch to 2-15/16 inches in length, numbering about 13,000.
In this combined rotor the steam thrust is aft for the ahead and forward for the astern rotor owing to the fact that steam is admitted at the ends of the casing and in each case flows towards the center or junction of the two systems. This central exhaust is shown in Fig. 2 and surrounds the astern casing. It should be particularly observed here that while the rotors are in one combined drum the casings are independent in a measure, the astern casing A, Fig. 2, being bolted on to the inner face of end of the main casing and being free to expand axially forward from that end. This modifies the lateral clearances of the blades in the astern rotor, for the entire rotor expands laterally from the forward end where the thrust collars are fitted.
As steam enters either one or the other end of this turbine according to which direction the engine is running, it is necessary to use some kind of packing to prevent it from by-passing the blades by entering the open ends of rotors. This packing is styled "Dummy Rings" C, and is a frictionless system of baffle rings nearly but not quite in contact. The details of these rings differ for ahead and reverse turbine owing to the necessity of providing for a greater range of expansion between rotor and casing at the far end.
Fig. 5 shows a section of both types—full size—giving not only a graphic idea of the shape and dimensions of the brass dummy rings (shown in solid black) but also of the clearance allowed. The astern "fin" type answers very well for backing period, but it is the ahead type which is most efficient and which demands so much attention as to its adjustment. In this the rings are set in groves in the casing and project into wider grooves turned in the dummy piston. The forward faces of these latter grooves are smoothly and accurately turned and the opposing faces of the projecting rings are hollowed out so as to reduce the surface to be "ground in" in initial fitting. The hollow also is the receiving face for leaking steam and tends to increase the baffling effect in eddying the current before it passes between the blade lips and collars.
There are 21 dummy rings in the ahead end of the L. P. turbine of the Preston but only 12 in the reverse end.
The means for keeping record of the clearance of ahead dummy rings are indicated also in Fig. 5. A brass stop piece D is fitted in the casing just forward of the rings, and against this the micrometer gage rod F (which enter casing through a stuffing box) is held by a spiral spring, in the outside details, when the rod is turned so as to bring its half diameter extension upwards. By turning the rod so as to bring this extension "finger" down it passes the stop piece and presses against the accurately turned and polished vertical edge of dummy piston and wheel BC. As the exact distance between the forward face of stop piece and the forward face of dummy piston is accurately recorded when originally set for proper clearance, any forward wear of rotor reduces this distance and the clearance by that amount and, by the necessary threaded sleeves and micrometer dials, this distance can be ascertained at any time while running, to within about one-thousandth of an inch; which Lieutenant Yates finds to be a common personal error in "feeling" the rotor and setting the gage.
The ahead turbines being the ones generally in use the closest possible adjustment of the dummy ring packing at the steam entrance end of these turbines is of absolute necessity for economical steam consumption. If the clearance is too great the steam loss is enormous and efficiency destroyed, while if the wear forward is allowed to continue until the rings bear, the brass strips may be ripped out of the grooves and a complete breakdown result by steam short circuit. In fact a clearance of two one-hundredths (.02") of an inch between the lips of the rings and the faces of the grooves as a running clearance when parts are hot should not be exceeded in the ahead dummy packing so it is obvious that this fine adjustment must call for special provisions in the details, and these are contained in the thrust bearings and shaft couplings.
In Fig. 2 it will be seen that the thrust bearing is placed forward instead of aft of the turbines, and in Fig. 3 it will also be noted that the shaft coupling forward (where it connects with another turbine shaft) is of the loose sleeve "gear" type which prevents axial thrust of forward turbine being transmitted to the after one. The object of this latter is to permit the dummy ring adjustment of either turbine independently, which could not be done if the couplings were solid.
The object of placing the thrust bearing at forward end and made part of the casting of each turbine casing is primarily to reduce as far as practicable the difference between the axial expansion of the rotor and that of the casing at the dummy packing, for it is obvious that this difference will increase with the distance from the end at which rotor and casing have a common fixed point. This fixed point can be considered to be at the middle of the thrust block where the thrust bearing is locked to the casing, as will be described, and with the short distance, as thus limited, it is practicable to control packing clearance very effectively.
In Fig. 2 the lower half of thrust block D is shown in position in the casing. This block rests on ribs of the casing as can be seen and is free to move axially within limits, as the lower half of thrust block has one rib, or shoulder, itself which lies between two of the ribs of casing. This can be noted in the photograph between the third and fourth bolts of this bearing, counting from the forward or near end.
It will be seen that the spaces between this shoulder of the thrust block and the ribs of casing are almost square and the locking of the thrust block in position is accomplished by filling these spaces with accurately turned half rings, of composition, after the rotor is adjusted to the desired position to give proper dummy ring clearance. By means of a simple packing yoke contrived to fit over a thrust collar on the rotor shaft, the rotor is screwed forward until the dummy rings and collars are in contact and then set back until the clearance is just exactly what has been determined upon as proper when cold. As the lower half of thrust block is for ahead thrust only, this part is now locked in position to hold the shaft from coming any further forward. The forward locking ring of course requiring the nicest fitting. The top half of thrust block takes the backing thrust only and is adjusted by means of side bolts on the cap. In this adjustment an axial play of from five to ten one-thousandths inch is permitted to the shaft between its bearing on go-ahead and astern thrust collars, as "float" for proper lubrication.
Whenever dummy packing clearance is to be re-adjusted new locking rings must be fitted in lower half of thrust block and daily, or very frequent measurement of the clearance is necessary for safeguarding the wear; but as accurate micrometer reading of the clearance can be taken while running this record is not very difficult to keep within one-thousandth of an inch.
While steam is kept from entering the interior of the rotors from the initial end by the above described packing, it is free to enter at the exhaust or discharge end and fill the rotor, so that provision must be made to prevent it blowing out along the shaft. This is also accomplished along the lines approaching frictionless packing, so important in turbine work.
On the rotor shaft between the spider and bearing B (Figs. 3 and 4) there is what is termed a Steam Gland, consisting of a set of fin type dummy strips and a set of loose brass collars (split and lap jointed) having an easy fit in grooves in the shaft but bearing tight up against the gland casing with their outer faces or rims. The casing also has corresponding fins to suit the dummy strips on the shaft and between the fins and the collars are spaces or pockets which are used either for connecting a "leak-off " pipe to carry away any leakage from a high pressure turbine to a lower or to admit low pressure steam to balance the steam leakage and to prevent air leakage into low pressure casing. This gland packing is seldom fully efficient and there is always some slight steam leakage into the engine rooms. A different form, and one which is reported to be working very well, is used on the Otaki (referred to later) and is simply soft packing in place of fins and collars, with the middle pockets kept filled with fresh water under pressure; but in this vessel no steam entered the interior of the rotor drum, it being completely closed at the ends, and hence the steam at the glands was only that which passed the dummy packing on rotor.
In the photograph of lower half of turbine casing it may be seen that there is but one moderate sized bearing at each end for supporting the shaft and rotor, and it is a notable fact that the downward wear is extremely small, in spite of the weight. Sometimes the wear is more upward, on the cap, than on the bottom brass, However, as the tip clearance of the blading of the turbine is maintained by these bearings it is necessary not only to provide perfect lubrication, but also to limit the thickness of the "babbit" metal or lining to not more than half that clearance so that should the bearing heat, and the white metal melt out, the shaft would be upheld by the brass ribbing in the box itself. The brasses are usually bored out only 30/1000 inch greater diameter than the white metal (or journal diameter) which leaves only 15/1000 inch of white metal on the brass between pockets.
These bearings are, in the case illustrated, about 13 feet apart and aside from the fact above noted of lack of downward wear experienced, there is another point of interest in the lack of sag. There does not appear to be any tendency to sag in these rotors—despite their weight and built-up construction—but their speed of rotation influences both of the features. In the L. P. rotor of the Preston the drum surface revolves at about 135 feet a second, or over 90 miles per hour.
In building the rotor the wrought steel drum is first bored out true and the outer surface left rough turned as received. The spiders, or "wheels," are usually of cast steel though those of the Preston were solid forged steel discs machined out to finished design to insure uniformity of material and balance. These are bored out for shaft ends and shrunk on same as no shaft passes through the drums, and then the spider wheels are trued up from shaft centers. In large spider wheels the shaft is further secured by large tap bolts half in shaft and half in wheel hub, axially, and riveted over into scores in bolt hole, making a flush joint. The "wheel" is now balanced before the drum is shrunk on and when both wheels and shaft ends are ready the drum is shrunk on the wheels and additionally secured by through tap bolts. In cases where the shaft is bored hollow a screw plug is fitted in the inner end, but in large work the boring out of shaft is not entirely through, but is for a distance from each end, leaving a solid portion midway.
After the drum is machined and bladed and the journals trued up, the blades are all turned off accurately as a final finish to them, this being very neatly and smoothly done by using a wide and straight edged cutting tool placed at a good "shaving" angle and drum revolved at very moderate speed. Not even a burr is left on the blade tips and uniformity of diameter for each step is secured.
The entire rotor is now finally balanced at speed of revolution required for actual full power. The shaft is mounted in bearings bored out to just permit easy turning of rotor, and these are set in pillow blocks an inch too wide; the ½ inch space on either side of bearing being filled with pure gum sheet; thus permitting the vibration of the shaft to be obvious and enabling the observer to readily note effect of balance weights. Speed is obtained by electric motors and sliding clutch connecting rotor shaft, which are arranged to be thrown out instantly when speed limit is reached. The shaft is chalked while spinning, for vibration, in the usual way, and trial weights clamped on in position indicated to be necessary. When the location and weight of the final balance piece is determined it is rivited to the arm as shown at F, Fig. 3. In the larger rotors the wheels are cast with thickening pads on inside of rim between each pair of arms, and these are chipped off as much as needed, in balancing, and afford a superior means of accomplishing this vitally important part of turbine building.
The balance secured is almost perfect and the test of standing a penny on edge on the casing over shaft bearings after complete assembling and while turbine is speeded under steam to full R. P. M. is commonly made without the coin's downfall.
The propellers are comparatively small—those of the Preston being but 5 feet in diameter. To any one unfamiliar with the changes brought about in adopting this type of engines, the statement that one wheel of that diameter is to transmit 3500 horsepower is surprising. The blades of these wheels are very highly polished and it is understood that in England they are sometimes silver plated. This is to be adopted by the New York Shipbuilding Company in future and undoubtedly is giving an unusual, yet most proper weight to the importance of reducing skin friction of high speed propellers to the very lowest degree possible.
In this article it is not proposed to go into any extended discussion of the turbine as a heat engine and the reasons for its superior economy at full powers ever the reciprocating engine. By the trials of the Michigan we are discovering greatly advanced economy in the latter type by using superheated steam and greater cylinder ratios. But no one who has studied the turbine question at all fails in noting the necessity of maintaining as nearly a perfect vacuum as is possible. The efficiency of the turbine lies mainly in its more thorough utilization of the heat in the low pressure steam than can be accomplished with reciprocating pistons and cylinders, mainly owing to loss in latter in cylinder condensation, and as Mr. J. W. Southern puts it in his excellent book, "The Marine Steam Turbine" . . . . the heat drop for a given pressure drop, increases as the pressure decreases and as the kinetic energy given up to the blades depends entirely upon the heat drop it naturally follows that in the case of very low pressure steam the same amount of work can be done with a much smaller pressure drop than with high pressure."
Cylinder condensation obviously is trifling in turbines compared to that in reciprocating engines as in the latter the alternate heating and cooling on pressure and exhaust strokes greatly varies the temperature of cylinder walls, which in L. P. cylinders must be corrected by jacketing and heat lost thereby. In any portion of a turbine cylinder the temperature must remain very uniform as the steam is in constant flow and fairly constant pressure at each step.
The arrangement of turbine machinery in the ship is a subject which should be most carefully thought out before we commit ourselves to general adoption of the new system. In the plans for the latest battleships a combination of reciprocating engines and low pressure turbines make a most enticing lay out (see Fig. 6) for we can have a four-shaft arrangement with the reciprocating engines on the wing shafts where large diameter of propellers are possible and the L. P. turbines on the middle or inboard shafts and there can be no question as to the resultant greatly increased economy, especially at low speeds, over all turbine.
The writer is not of the opinion that the combination of the two types is altogether desirable. No doubt it is more elastic and has economical advantages in coal consumption at low speeds, but we are, in using this, clinging to the older type of machinery and retaining all the troubles of their adjustment and upkeep; while the real step toward progress is to give up the reciprocating engine and adopt rotary type as soon as we can possibly do so without too great sacrifice.
No one who has not had long experience with reciprocating machinery, can appreciate the relief one feels at being rid of the adjustment proposition of bearings in this class and the proper care of all the connections, piston rings and joints. The forever watchfulness necessary for lubrication and the anxiety and worry due to hammering at crankpin journals after days of fast steaming, all cry for riddance of the type so long our stand-by and, oddly enough, just leaping forward, economically, through the use of superheated steam and changed proportions.
The Otaki, the first steamer using combined turbines and reciprocating engine, after making several voyages between England and New Zealand, seems to point to a very economical arrangement showing even at twelve knots a coal consumption of only 1.387 pounds of coal per horse-power per hour; but this is an arrangement of two triple expansion reciprocating engines and only one (middle) L. P. turbine. An interesting account of this machinery and performance can be seen in the London Engineering of August 6, 1909, and is worth reading.
As to weights, we have no uncertain reduction by using all turbines for any power. The London Engineering of August 13, 1909. states it as follows:
INDICATED HORSE-POWER PER TON OF TOTAL MACHINERY WEIGHT.
Type of Ship. Reciprocating Parsons’
Engines. Turbines.
Large armored warships 12.00 14.00
Protected cruisers, scouts, etc. 19.00 25.00
Torpedo-boat destroyers 45.00 65.00
Cross-Channel steamers 9.00 12.00
Passenger liners 7.00 …
In our present battleship installations we are obliged to govern the speed of all four shafts with one throttle when under any but full speed conditions, where we have one throttle to govern each pair of shafts, and this has already proved to be a serious drawback in maneuvering. To illustrate, Fig. 7 shows an outline layout of the turbines of the Utah. Here we have, for full speed, steam admitted separately to each main high pressure turbine and thence to the main low pressure on the next shaft so that the speed of two shafts only is controlled by one throttle, and the course of the steam is indicated by the lines marked A1, A2
For the lowest cruising speed, indicated by lines B, we have to admit steam first to the H. P. cruising turbine and from this it goes to the M. P. cruising turbine and thence to both main H. P. and both main L. P. in sequence, the H. P. cruising throttle being the regulator.
For intermediate speeds we cut out the H. P. cruising turbine altogether and begin with M. P. cruising as indicated by lines C and from there to both main H. P. and L. P. turbines as before, the M. P. cruising throttle governing all shafts.
Economy in the turbine depends upon the relation of the speed of the blades with that of the steam flow due to the difference in initial and final pressures in each step. The blades should travel half as fast as the steam flow for best work. It is impracticable to attain this perfectly without very large diameters or very excessive R. P. M. so it is to be approached in compromise. .In full speed we can approach it to the high R. P. M. utilized, in simply the main H. P. and L. P. turbines, but when we come to reducing the R. P. M's. we have to use a greater number of steps to reduce the steam flow in each to something near the theoretical.
It is more than doubtful, however, that economy gained by such extended complications is worth the cost of installation and detriment to maneuvering facilities and it is not surprising to note that the British Navy is cutting out some of this in later installations and simplifying the arrangement so as to have at least one governing valve for each pair of shafts.
In Fig. 8 is shown a layout that accomplishes as much as should be needed, especially as cruising speeds are sure to increase from what they were or are, to somewhere around 15 knots. Here we have simply a H. P. cruising turbine added to and forward of the main H. P. turbines, with suitable by-pass piping, so that when we desire to run at slower speeds we give steam first to the cruising turbines and thence to the main H. P. and L. P. as indicated by the lines B.
The immense saving of space in an important part of the engine rooms is obvious, and the maneuvering facility is much greater than in Fig. 6, and while the economy is possibly less at lowest speed, a reduced initial pressure at the boilers would make up in a large measure even for that.
We can feel assured that the marine turbine has come to stay and that our own engineers will see to it that it is improved more rapidly in the next ten years than it has been since first installed in a seagoing vessel. Numerous changes for its betterment are already afoot and result from a still very limited practical experience, but with the haste necessary to complete a vessel on contract time, there is little possibility of avoiding imperfections or of incorporating novel features occurring to the builders during the work. Each successive plan will be an improvement on its predecessor and gradually the new will be made excellent.