KITE METEOROGRAPH, CONSTRUCTION, AND OPERATION.

By Prof. C. F. MARVIN.

In designing the kites, meteorograph, and aerial apparatus employed in making the observations discussed in the accompanying paper, it was necessary to satisfy a number of conditions and limitations, as, for example, portability, simplicity, and strength of construction, especially as regards the kite; uniformity and interchangeability not only of parts of the apparatus, but of parts of the same devices, so that repairs and renewals could be made as easily as possible. The corps of observers, moreover, when they began, were without previous experience in kite flying. The entire class, however, reported at Washington before the individual members were assigned to station, and was put through a preliminary drill and course of instruction in the practical work. In the short period of ten to twenty days devoted to this instruction (including the time lost during unfavorable weather) it was possible for the observers to gather only a general knowledge of the apparatus and methods of work.

The kite. In conformity with the foregoing limitations, it was considered best to employ in each ascension only one kite, rather than several in tandem, which latter are very troublesome to set flying in light and fitful winds and are not required; in fact, are less efficient than a single kite in favorable winds.

Nearly all observations were made with a medium size kite containing 68 square feet of supporting surface. At some stations a smaller and a slightly larger kite were also sometimes used, according to the strength of the wind. The smaller size contained 45 square feet and the larger 72 square feet. Fig. 1 shows the kite with the meteorograph in place. The dimensions of the medium size kite are as follows:

Transverse width of kite..

Length, over extreme edges, fore and aft1

Distance between top and bottom supporting surfaces
Width of cloth bands

6 feet 6 inches.

6 feet 2 inches.

2 feet 8 inches.

2 feet.

These kites are framed and constructed in the most rigid manner possible. The six longitudinal sticks running fore and aft are attached by means of small machine screw bolts to the rectangular frames forming the rigid edges to the cloth bands and are detachable; thus permitting the kite to be collapsed, as shown in fig. 2.

The best cloth material for kites seems to be Lonsdale cambric, which is light, strong, and closely woven. A black, or dark-colored cloth is more visible on many occasions, and the rear cell was covered with black nainsook on this account. This cloth is not as strong as the cambric, but, owing to the circumstance that the pressure per unit area on kites of this type is very much greater in the forward than in the rear cell, the above disposition of the relatively strong and weak material is wholly justified. This same circumstance explains why the front cell is made with three supporting surfaces, as compared with two in the rear cell; that is to say, this increase of surface (at slight increase of weight) increases the lifting efficiency of the kite.

The length, fore and aft, was increased to 6 feet 8 inches, in later forms.

7

The flying line is attached to the front edge of the forward cell at the middle; the ordinary arrangement of this connection being more fully shown in fig. 3.

Bridle and safety line. The normal bridle, including the safety line for the kite, is shown in fig. 3. The cord (No. 32 Italian blocking cord) at A is passed twice around the stick, after the fashion of a "clove hitch;" the free portions of the line, one of which is rather short, are firmly bound together where they emerge from the clove hitch by a serving of waxed "gilling" thread. The long end is passed through one of the metal safety-line eyes, B, and the two ends tied together by means of bowline knots, as at b. A similar, but longer, piece of cord is secured at by a clove hitch and the free ends attached to each other by bowline knots after the long end has been passed through the eye B'. When B' is held so that the line AB' is taut and at an angle of 900 to the stick, the line C'B should be just a little taut. The main line is attached to the bridle at B'. Safety line. The eyes, BB', are connected by what is termed a safety line, S, which is simply a piece of steel wire, the size of which is so chosen that its ultimate strength is within a safe working strain for the kite and flying line. In normal flight all the strain produced by the kite upon the line is transmitted through the safety line S. If in any case the conditions give rise to a greater strain than the ultimate strength of the safety line (and therefore dangerous to either the kite or the line, or both), it will be broken. In this case the kite will thereafter fly from the point C, which is several inches in advance of A. Such a change in the bridle causes a diminished pull by the kite, other things being the same.

The safety lines generally supplied have a tensile strength of about 85 pounds.

The reel. The management of large kites in flight requires a substantial and convenient form of reel of the character indicated in fig. 4. The top portion of the carriage revolves upon the table below on bearings resembling the so-called "fifth wheel" of wagons. The drum revolves easily in metal bearings and is fitted with dials at the axle indicating the number of revolutions. At the start the dials stand at zero and count off revolutions as the wire unwinds.

Strap brake. The lever seen at the right in fig. 4 operates a powerful strap-iron friction brake acting on the rim of the drum and controls in the easiest and most complete manner the unwinding of the wire or the stoppage of the reel under all circumstances.

A matter of great importance in the design of the winding drum of the reel is to secure sufficient strength in the rim to withstand the enormous cumulative pressure exerted by a large amount of wire wound in under great tension. A single turn of wire around the drum, under a uniform strain of 50 pounds, for example, tends to produce a compressive stress of 50 pounds at every point around the rim. The next turn, at the same tension, adds 50 pounds to the preceding stress, and so

on.

Two thousand turns at this rate will, therefore, produce a pressure of 100,000 pounds, or 500 tons. The heavy rim of the cast-iron drum, shown in fig. 4, is calculated to safely resist a crushing pressure of fully 1,000 tons. In actual practice the crushing pressure is not quite so great as that calculated by the process indicated above, because the material of the reel yields a little as the pressure increases, and this lessens the tension on the turns of wire already wound on the drum. The side flanges of the drum must also be very strong, as the wire crowds sidewise against these with great force. It is best on this account not to wind the wire on in smooth and even layers, but rather to crisscross the turns of wire slightly, but in a regular manner. Wound in this way, the wire tends to support itself, even without side flanges; at any rate, the lateral pressure is greatly reduced, and, moreover, the outside turns of wire are not able to squeeze down through what is already wound on the reel, as they tend to do when the wire is wound in an even manner, like thread on a spool.

When flying at an elevation of from 5,000 to 7,000 feet, one of the Weather Bureau kites, supporting its instrument, will pull from 60 to 80 pounds, if not more, and from 8,000 to 10,000 feet of wire will be out. To wind all this wire in under such conditions is really a very laborious operation, and generally requires two men at pretty hard work for from a half to three-quarters of an hour or more.

As sent out to stations the hand reels contained from 2,600 to 3,000 turns of tempered steel music wire, 0.028 of an inch in diameter. The normal tensile strength of this wire was about 200 pounds.

Length of wire. As the original supply of wire was wound upon each reel, record was kept of

the total number of turns and a table computed, giving the number of turns corresponding to given lengths of wire in units of 500 feet. Due account is taken in these tables of the gradual diminution in the length of each turn as more and more wire is unwound. The coefficient of diminution was determined from several sets of readings of the revolutions of a measuring wheel around which the wire passed as it was being wound on a reel. Simultaneous readings of the dial on the reel were also made. The measuring wheel was accurately 3 feet in circumference and the dial indicated feet.

Electrical connections.-The wire line employed in flying kites becomes electrified more or less at all times, often highly so. For comfort of the operators, as well as safety, this charge must be conveyed to earth, and for this purpose each reel stand is provided with an electric groundconnection and switch. The 15-inch cranks by which the reel is revolved are made of wood for the sake of insulation.

Radius rod and arc.—It is important to know the inclination of the wire at the reel in order to make a proper allowance for the sag of the wire. This is accomplished by means of the radius rod and the graduated arc, fig. 4. The radius rod is clasped loosely upon the axle of the reel on either side of the drum, and the arc hung over the shaft on a pair of antifriction wheels which run in a groove turned in the shaft. A weighted rod below the arc causes it to maintain a vertical position at all times, thus insuring correct angles. In use, the radius rod is made to rest. against the wire, the angular inclination of which to the horizontal is then shown by the reading upon the graduated arc. This angle is subject to a small and variable inaccuracy, due to slight alterations in the radial distance of the wire at the point it leaves the drum, according as more or less of the wire is unwound.

Dynamometer. The tension upon the line at the reel at any time is determined by means of the dynamometer permanently attached to one of the crank handles, as seen in fig. 4. This consists of a short, stiff, steel spring, firmly fastened to the outer end of the handle at the back. The short end of a multiplying lever connects with the spring, while the long end serves as an index and traverses the graduated arc shown on the crank handle near the axis. The reading on this graduated arc indicates the pull in pounds on the wire. Here, likewise, a small and variable error is introduced because of the variations in the diameter of the drum with different amounts of wire out. This is not important.

End of wire. The outer end of the wire terminates in a small brass eye. To facilitate connecting the wire to the kite, a piece of No. 32 blocking cord about 8 feet long is fastened to the eye-that is to say, the cord is simply passed through the eye and a bowline knot tied on one end.

Reel box, cover, and lock.-When not in use it is designed simply to inclose the reel within the box by means of a suitable cover. The crank handles are unshipped and placed inside the box. The cover is secured and locked by means of an eyebar, which passes between the spokes of the drum and is fastened by a padlock on one end.

The meteorograph.-The instrument sent up with the kite to secure the automatic record of the conditions of the air is seen in fig. 1 as it appears attached to the kite and inclosed within its light aluminum case. The mechanisms inside the case are shown in fig. 5, and are designed to record wind velocity, temperature, pressure, and humidity of the air. Records of the velocity of the wind were, however, made experimentally only at Washington, and were not included in the official observations.

Special consideration was given in designing the meteorograph to secure a proper exposure of the hair hygrometer and the thermograph bulb. The former consists of a strand of prepared hairs stretched back and forth in a double strand and nearly from end to end inside the long tube seen at the top portion of the instrument, as shown in fig. 5. The direct elongation and contraction of these hairs is communicated to the recording pen at the extreme right side in fig. 5. The thermograph bulb is also placed within the long tube and occupies nearly a middle position.

When attached to the kite, as shown in fig. 1, the meteorograph is so placed that the wind blows with full force directly through the tube containing the thermometer bulb and hygrometer, thus affording thorough and complete ventilation, radiation being at the same time effectually cut off. Even though the metallic case of the instrument becomes heated on exposure to sunshine, yet the

metal tube itself is not only shaded and exposed to a strong current of wind, but is everywhere separated from contact with the metallic case by vulcanite rings at the ends and longitudinal ivory strips on the sides. Indeed, celluloid strips are interposed between the thermometer bulbs and the metal of the inclosing tube, thereby still further insulating the thermometer bulbs. These bulbs consist of a pair of tempered-steel bourdon pressure tubes, forming a curl of about seveneighths of a complete circle about 1 inches in diameter. The major and minor axes of the elliptical cross section of the tubes measure approximately 0.5 and 0.1 of an inch, respectively. The tubes are filled with pure alcohol under pressure, and are set in tandem and edgewise to the current of wind through the tube. Thus arranged, they constitute highly sensitive and relatively powerful thermometric bulbs. The recording pen traverses a scale of 22 degrees to the inch, each instrument being adjusted to this scale by tests at different temperatures, the air being driven through the tube by means of an electric fan. The record sheet provides a range of 45 degrees possible change of temperature in any ascension, the initial setting of the thermograph pen being effected by a suitable adjusting screw.

The mechanism of the aneroid barometer is readily understood from fig. 5. The novel feature in this part of the meteorograph consists in the use of tempered and highly elastic steel corrugated disks, instead of brass or German silver ordinarily used. Numerous tests showed the steel vacuum chambers to be superior to others. Nevertheless the aneroid principle has not thus far proved to give pressure records characterized by that high degree of precision required in meteorological work.

The pressure scale on the record sheet embraces a range of 9 inches, viz, 21 to 30 inches air pressure. The subdivisions of the sheet are 22 spaces per inch, each space representing 0.2 of an inch barometric pressure.

It will be noticed these spaces are the same as the degree spaces on the temperature record. This arrangement was chosen so that in extreme cases the temperature record might overlap on the space normally provided for pressure, and vice versa, and still be properly scaled by the rulings on the sheet. Furthermore, in designing the instrument the scales were so chosen that in an ascension under average atmospheric conditions the pressure and temperature curves would have about the same actual amplitude. Hence the change of temperature in a given elevation is measured with the same accuracy as the corresponding change in pressure, at least so far as the traces themselves are concerned. A greater precision of pressure measurement is desirable, but is scarcely justified by the inherent defects of the aneroid principle of measurement.

Each recording pen is adjustable, and at or before the time of ascension the pens are set as nearly as may be to indicate correctly the atmospheric conditions as shown by readings of the sling psychrometer and mercurial barometer. After these settings are made and just before ascensions, the meteorograph having been exposed for some time in the wind, readings of the psychrometer and barometer are again made and the outstanding differences of the indications of the meteorograph noted and used as corrections to the automatic records, similar corrections being noted at the close of the ascension.

Elevation of the kite.-The vertical height of the kite and hence the meteorograph above ground was determined from readings of the angular elevation of the kite, the length of line out, and its inclination at the reel. The angular elevation was measured to the nearest half degree by means of the nephoscope, shown in fig. 6. The position of the sighting staff also indicated the angular azimuth of the kite.

The approximate elevation of the kite was taken out from a table giving values of the expression hl sin q. A percentage correction was then applied, depending upon the amount of sag in the wire as shown by its observed inclination at the reel.

Table I gives the percentage of slack in the wire, as deduced from the equation of the catenary for such conditions as commonly occur.

Table II is a copy of one of the cards furnished with reel No. 15, giving the number of turns of the reel corresponding to the lengths of wire out in 500-foot intervals.

As a rule, observations were made with whole units of 500 feet of wire out, so that interpolation was not necessary. The length of wire in one turn was, however, accurately tabulated, and facilitated accurate interpolation.