CHAPTER VI

CONSTRUCTION OF SUPPORTING SURFACES

An examination of the wings of birds, whether those of soarers or of any other type, impresses one not only with the general strength of the wing, but also with the fact that, while it possesses considerable stiffness, there is also a graduated pliability, not only of the whole wing, including the bones, but more especially in the feathers, the rear tips being exceedingly pliable so that, when the wing is held in a stiff breeze, they are seen to be easily deflected in a gentle curve towards the rear and upper side. This lack of rigidity has several advantages, among the more notable of which is the lessening of the strains on the wing caused by sudden wind gusts. Of great importance is the further fact that a supporting surface having a graduated pliability, such as is possessed by a bird’s wing, does not experience a shifting of the center of pressure to the same extent as a rigid surface of similar form. Furthermore, since any bird, even the best soarer, must use its wings not only for soaring, but, when starting to fly from a state of rest, for flapping, a rigid surface would not furnish anything like the same universally available sustaining and propelling means that the bird’s wing does.

In an inspection of the various wings or supporting surfaces which Mr. Langley built, from the very earliest rubber-pull models up to the successful steam machines Nos. 5 and 6, the point which is most impressed upon the observer is the increasing strength and rigidity embodied in these wings. While the success with the later models was due to many things, including the development of a strong frame and a suitable power plant giving sufficient power for the permissible weight, besides the very important development of effective equilibrium mechanism, yet it is safe to say that even with the development of all these other things to the state to which they had been brought in 1896, success would not have been achieved had not the wings themselves been simultaneously changed from the very flimsy construction which was at first used to the later type, using a very strong and rigid wooden frame over which the cloth covering was tightly stretched, and which possessed only a small amount of pliability at the extreme rear ends of the cross-ribs.

The development of this successful type of wing for the models, it will be remembered, had been achieved only after an extensive series of experiments; and it was realized that the construction of suitable wings for the large aerodrome, even with the knowledge gained in the early work, would be still more [p189] difficult. The problem was that of constructing for a very little greater weight per square foot, wings containing approximately sixteen times the area of the model wings.

It will be recalled from the previous description of model Aerodrome No. 5, that its four wings had a combined area of 68 square feet and weighed approximately 2500 grammes, or 37 grammes per square foot. It was not expected that the large wings would be of so light a weight per square foot, which would have meant only about 35,500 grammes (approximately 78 pounds) weight for the 960 square feet originally planned. It was hoped, however, that the increase in weight per square foot for the large wings would be less than the square root of the increased linear dimensions. In this case, the increase in linear dimensions being approximately four, it was, therefore, hoped that the larger wings would not have quite twice the weight per square foot of the smaller ones; the computed weight permissible for the large wings was therefore placed at 120 pounds.

To obtain the required area within the permissible limits of weight two well-defined paths of procedure were open: First, it was possible to so modify the structural form of the wing as to obtain the advantage of the increased strength of trussed structures, that is, by superposing the wings. Or, second, the “single-tier” type of wing, the efficiency of which had been fairly well determined, could be retained, and strength gained without increase of weight by improving the method of constructing the wooden framework and by extending the system of guy-wires.

Some knowledge of the superposed type of supporting surfaces had already been gained by the experiments at Allegheny and the tests of the rubber-driven models, in which superposed wings had frequently been used; but it was felt that this knowledge was altogether inadequate to aid in determining either whether the superposed type of construction possessed in practice the advantages which theory would indicate, or how and at what distance apart the surfaces should be superposed to obtain the best results. In order to obtain the desired information, a series of tests on the whirling-table of complete wings suitable for use on the models was made. These experiments were supplemented by the practical tests with the models, which have already been described in Chapter III, in order to give the wings a trial under the conditions of flight, where they would be subjected to the action of the propellers and the uneven character of the wind.

In addition to determining what type of construction and what form of surface would give the greatest “lift” with the smallest “drift,” these whirling-table tests supplied data as to how much greater the actual resistance of the wing with its necessary guy-posts and guy-wires was than the theoretical resistance, found by extrapolation from the results obtained in the tests of rigid [p190] curved surfaces formed of wood. The first of this series of tests, the results of which are given below, was made November 30, 1898, on the superposed wing shown in Plate 37, Figs. 1 and 2. It should be noted, however, that when this test was made the wing was not provided with the stiffening strips or the vertical partitions.

Weight of wing = 1000 grammes; weight of guy-posts, etc., = 475 grammes; distance of mean center of gravity of guy-posts, etc., from pivots of balance arm = one-half distance of CP of wing from pivots of balance arm; the wing, therefore, had a lever arm of two to one with reference to weight of guy-posts, etc., so that the equivalent weight of guy-posts, etc., = 237 grammes. This gives 1237 grammes of equivalent load on the wing = 2.73 pounds. Area of wing = 21.85 square feet. Therefore load on wing = 0.125 pounds per square foot.

Angle Revolutions Velocity of Velocity Drift of of turn- center of (ft. per (grammes). chord. table. wing (ft. second). per min.) 2.0° 10.75 1086 18.1 255 3.0° 10.0 1010 16.85 255 5.0° 9.5 960 16.0 255 10.0° 7.75 783 13.0 255

Angle Drift Foot-pounds Calculated of (pounds). per sec. soaring speed chord. RV. carrying 0.5 pounds per sq. ft. (ft. per sec.). 2.0° 0.561 10.15 36.2 3.0° 0.561 9.47 33.7 5.0° 0.561 8.98 32.0 10.0° 0.561 7.3 26.0

The very interesting phenomenon was noted in this test that the “drift” or resistance of the wing seemed to remain unchanged at soaring speed at different angles of elevation. It is hardly probable that this result is accurate, for the “balance arm” undoubtedly twisted under the action of the wing, and this caused it to strain on its pivots, and thus, to a certain extent, falsify the record as to drift.

A test of a single-tier wing at different angles of elevation was made on December 6, 1898. This wing was nearly the same as those used in actual flights of Aerodromes Nos. 5 and 6 in May and November, 1896, the wing being of the same width fore and aft, but somewhat shorter. The actual wing was a little too long to permit its being used on the whirling-table in the limited space of the shop.

Weight of wing = 420 grammes; weight of guy-posts, etc., = 320 grammes; equivalent weight of guy-posts, etc., = 150 grammes applied on the wing. Therefore, total load on wing = 570 grammes. Area of wing = 11.2 square feet; equivalent load on wing = 0.112 pounds per square foot.

Angle Revolutions Velocity of Velocity Drift of of turn- center (ft. per (grammes). chord. table. of wing second). (ft. per min.) 2.0° 11.6 1195 19.9 210 3.0° 9.75 1005 16.7 157 5.0° 8.25 850 14.2 133 10.0° 6.75 695 11.6 129 12.5° 6.0 618 10.3 129

Angle Drift Foot-pounds Calculated of (pounds). per sec. soaring speed chord. RV. carrying 0.5 pounds per sq. ft. (ft. per sec.). 2.0° 0.462 9.2 42.1 3.0° 0.345 5.77 35.3 5.0° 0.293 4.16 30.0 10.0° 0.284 3.29 24.5 12.5° 0.284 2.92 21.8

In this test it is to be noted that the “drift,” or resistance, while considerably greater at soaring speed for 2 degrees than for 5 degrees, remains practically the same between 5 degrees and 12-1/2 degrees. Comparing it with the preceding test with the superposed wing, it is seen that at soaring speed at an angle of 10 degrees, the single-tier wing having a load of 0.112 pounds per [p191] square foot, has only 129 grammes drift, while the superposed one, while supporting 0.125 pounds per square foot, has 255 grammes drift. Moreover, the soaring speed of the single-tier wing is only 11.6 feet per second, while the superposed one requires a speed of 13 feet per second.

As the superposed wing tested on November 30 was so weak structurally that it could not be made to keep its proper shape without adding an excessive number of guy-wires, it was decided that it was not adapted for use on the aerodrome, but before abandoning it the partitions and strips were added and it was again tested on the whirling-table on March 1, 1899, with the following results:

Weight of wing = 905 grammes; weight of guy-posts, etc., = 320 grammes; equivalent weight of guy-posts, etc., = 150 grammes applied at ‹CP› of the wing; equivalent load on the wing = 1055 grammes = 2.321 pounds; area of wing = 21.85 square feet; equivalent load on wing = 0.1062 pounds per square foot.

Angle Revolutions Velocity of Velocity Drift of of turn- center (ft. per (grammes). chord. table. of wing second). (ft. per min.) 5.0° 10.875 1100 18.35 250 5.0° 10.75 1085 18.07 250 5.0° 10.75 1085 18.07 250 10.0° 8.0 808 13.47 250 10.0° 8.0 808 13.47 250 10.5° 7.875 797 13.3 250 10.5° 7.875 797 13.3 250 13.0° 7.0 707 11.78 250

Angle Drift Foot-pounds Calculated of (pounds). per sec. soaring speed chord. RV. carrying 0.5 pounds per sq. ft. (ft. per sec.). 5.0° 0.55 10.1 39.81 5.0° 0.55 9.94 39.19 5.0° 0.55 9.94 39.19 10.0° 0.55 7.4 29.226 10.0° 0.55 7.4 29.226 10.5° 0.55 7.32 28.86 10.5° 0.55 7.32 28.86 13.0° 0.55 6.48 25.553

An examination of the data obtained in this test shows the wing to be of slightly less efficiency than when first tested. While it was considerably stronger it was still too weak for use on the aerodromes.

A second type of superposed wing, Plates 64 and 65, was therefore constructed and tested on the whirling-table on March 2, 1899, with the following results:

Weight of wing = 1025 grammes; weight of guy-posts, etc., = 320 grammes; equivalent weight of guy-posts, etc., = 150 grammes applied at ‹CP› of the wing; equivalent load on the wing = 1175 grammes = 2.585 pounds; area of wing = 21.85 square feet; equivalent load on wing = .1183 pounds per square foot.

Angle Revolutions Velocity of Velocity Drift of of turn- center (ft. per (grammes). chord. table. of wing second). (ft. per min.) 5.0° 11.625 1170 19.5 250 5.0° 11.625 1170 19.5 250 8.0° 10.5 1060 17.7 250 10.0° 9.125 919 15.3 250 10.0° 9.125 919 15.3 250

Angle Drift Foot-pounds Calculated of (pounds). per sec. soaring speed chord. RV. carrying 0.5 pounds per sq. ft. (ft. per sec.). 5.0° 0.55 10.72 40.087 5.0° 0.55 10.72 40.087 8.0° 0.55 9.75 36.37 10.0° 0.55 8.43 31.4 10.0° 0.55 8.43 31.4

During the tests on the whirling-table this type of construction seemed to be exceedingly strong and stiff, and to be easily maintained in whatever position it was placed. It was therefore thought that it would prove strong enough for the aerodrome, and it was accordingly inverted and given a “sanding test” [p192] by sprinkling sand uniformly over it to such a thickness as to cause it to have a load of 0.75 pounds per square foot. As it showed no serious deflection or change of form under the sanding test, it was decided that it was strong enough for use in tests of the model aerodromes in actual flight.

Upon the completion of these whirling-table tests, the cloth covering of this wing was painted with collodion varnish, which increased the weight of the wing only 50 grammes. In order to make the results of its tests more easily comparable with those obtained before varnishing, the cross guy-wires on the wing were changed to a slightly smaller size in order to make the weight of the wing the same as before. It was tested on March 3, and the following results were obtained:

Weight of wing = 1025 grammes; weight of guy posts, etc., = 320 grammes; equivalent weight of guy-posts, etc., = 150 grammes applied at ‹CP› of the wing; equivalent load on wing = 1175 grammes = 2.585 pounds; area of wing = 21.85 square feet; equivalent load on wing = .1183 pounds per square foot.

Angle Revolutions Velocity of Velocity Drift of of turn- center (ft. per (grammes). chord. table. of wing second). (ft. per min.) 5.0° 10.5 1060 17.7 250 5.0° 10.5 1060 17.7 250 10.0° 8.5 859 14.3 250 10.0° 8.5 859 14.3 250

Angle Drift Foot-pounds Calculated of (pounds). per sec. soaring speed chord. RV. carrying 0.5 pounds per sq. ft. (ft. per sec.). 5.0° 0.55 9.75 36.37 5.0° 0.55 9.75 36.37 10.0° 0.55 7.88 29.4 10.0° 0.55 7.88 29.4

Although the varnishing of the wing seemed to have no effect on the “drift,” the soaring speed was slightly decreased.

As a result of these tests it was decided to construct three more wings like this second type, the four forming a complete set for use on the steam-driven models Nos. 5 and 6. Although the tests on the whirling-table indicated a superior efficiency for the “single-tier” wings, and it was not expected that in actual use on the aerodrome the result would be different, yet it was felt that as the conditions of actual use are so very different from those of a whirling-table experiment it would not be safe to decide too definitely against the superposed wings without first giving them a test under actual conditions. Aside from the decreased lifting effect shown by the superposed wing when compared with the “single-tier” one, it was also thought that under the actual conditions of use on the machine the superposed wing would show up still worse. The deflection of the air by the front wings diminishes the lift of the rear ones even for the “single-tier” type, and this, it seemed certain, would be greatly aggravated in the case of the superposed type.

In order to emphasize more fully the results of these tests the following table is added, which gives the data for the “single-tier” wing and this second type of superposed one, when each was tested at ten degrees angle of elevation:

[p193]

Wing “Single-tier” Superposed. Type No. 2 Length (feet). 4.27 4.27 Area (sq. ft.). 11.2 21.9 Weight (pounds). 1.26 2.59 Angle of chord. (°) 10 10 Soaringspeed, (ft. per sec.). 11.6 14.3 Weight (lbs. per sq. ft.). .112 .118 Drift (lbs. per sq. ft.). .025 .026 Calculated soaring speed, carrying 0.5 lbs. per sq. ft. (ft. per sec.). 24.5 29.4

The “single-tier” wings actually used on Aerodrome No. 6 were 5.33 feet long, while the wing tested above was only 4.27 feet long. In order to bring out more fully what might be expected of Aerodrome No. 6, when using the two different types of wings, the following table, calculated from the preceding one, is given. This shows the results which might be expected from the aerodrome when the resistance of the machine itself was included:

Aerodrome No. 6 without wings weighs 22 pounds.

“Single-tier” “Single-tier” Superposed. (short) (full-length) Type No. 2 Length (feet). 4.27 5.30 4.27 Area of two pair of wings (sq. feet). 44.8 54.0 87.6 Weight of two pair of wings (pounds). 5.04 5.5 10.36 Weight of aerodrome with two pair of wings (lbs.). 27.04 27.5 32.36 Total weight to be supported (lbs. per sq. ft.). 0.603 0.51 0.369 Drift of wings (lbs.). 6.06 6.13 6.9 Assumed drift of aerodrome body (lbs.). 1.0 1.0 1.0 Total drift (pounds). 7.06 7.13 7.9 Soaring speed (ft. per sec.). 27.0 24.7 25.3 Thrust horse-power expended. 0.35 0.32 0.364 Brake horse-power expended. 0.70 0.64 0.73

The first line shows the calculations for the aerodrome when equipped with the short “single-tier” wings; the second line, when equipped with the “single-tier” wings of the full length used in the flights of 1896; and the third line, when equipped with superposed wings, Type No. 2.

It will be seen that, on the whole, the result of the comparison of the full-length “single-tier” wing and the superposed one is less in favor of the latter than was to be expected, as, aside from its greater structural strength, it seems to have no real point of superiority, except that it is shorter; and, as already pointed out, one point of presumable inferiority, though not exhibited in the table, is the fact that the rear set of wings would suffer relatively more from being in the lee of the front ones, in the case of the superposed wings, than in the case of the “single-tier” ones.

Besides these “conventional” forms of wings, various other types were tested on the whirling-table. The data of these tests are not given, as in the rough preliminary tests the results were so entirely negative in character that accurate quantitative tests were never made. However, since in work of this kind the greatest delay is experienced in learning what not to do, and in ridding one’s self of freak notions which are continually suggesting themselves, it may be well here to describe sufficiently at least one of these types of wing to enable others to avoid any loss of time in experiments with it. Since the principal disadvantages of a wing possessing considerable width in the fore and aft direction are due to the great extent through which the center of pressure [p194] shifts when the velocity of advance or angle of incidence is changed, and to the further fact that a wide surface does not support proportionately as much per square foot as a long and narrow one, it was thought that some advantage might be gained by making the covering of the wing in the form of strips, the edges of which would be perpendicular to the direction of motion, or by making this covering in more or less slat-like form, which would permit the air which had already been acted upon by the leading slat to slip through between the rear edge of the first slat and the leading edge of the succeeding one. In the tests on the whirling-table, however, it was found that this type of construction not only did not possess any advantages, but was even less effective than a similar one in which the covering was continuous. The difference was probably due to the fact that the air which passed between the slats reduced the suction on the upper side of the following slat, and also to the fact that the distance between the slats was not sufficient to gain the effect of having each slat act on air which had not already been partially deflected by the preceding one.

In view of the results of these tests on various types of wings, it was decided that in constructing the first set of wings for the large aerodrome it would be best to employ the “single-tier” type, which had proved successful with the models, and that after getting a successful flight with these the superposed wings would be tried in order to get, if possible, the advantage which they possessed of being structurally stronger and more compact. It was therefore clear that any gain in the strength and rigidity of the first set of wings, as a whole, would have to be obtained by improvements in the construction of its integral parts, that is, in the main and cross-ribs which made up its framework.

Before attempting to proportion the parts of the necessary wooden wing frame, which it was expected would probably undergo many changes before a final design was secured which would embody maximum strength for minimum weight, various tests were made to determine just how light a cloth covering could be obtained which would be strong enough and sufficiently impervious to the air. In the construction of the wings for the models a good grade of China silk had been employed, but on account of the greatly increased quantity of cloth required for the large wings, it was hoped that something approximately as good as the silk could be secured at a much less cost, and various grades of percaline were therefore tested. The weight of the various grades of percaline ranged from three grammes to ten grammes per square foot, the lighter samples being of a rather coarse mesh, while the heaviest ones were not only close mesh but some specimens contained a large amount of “sizing.” The particular grade which was finally adopted weighed seven grammes per square foot. This material was practically impervious to air at a pressure of one pound per square foot, which, of course, was considerably [p195] more than it would be subjected to in flight. This grade of percaline weighed approximately one and a half times as much as a grade of silk, which on test was found to have a slightly greater tensile strength than the percaline, though the latter did not “flute” or “pocket” nearly as much as the silk. Moreover, the cost of the percaline was only about one-third that of the silk, and it was chiefly for this reason that percaline was adopted in place of silk. Allowing for necessary seams and extra material to be turned over at the front and rear edges of the wings, the percaline covering, which under the original plans comprised approximately 1000 square feet, was therefore calculated to weigh approximately 7000 grammes, exclusive of the necessary cords for lacing the coverings to the wooden frames of the wings.

As the one hundred and twenty pounds allowed for the four wings permitted only thirty pounds per wing, and as the cloth covering, lacing cords, etc., were found to weigh something over four pounds, there remained only about 25 pounds as the permissible weight of the wooden framing, including the necessary metal clips, secondary guy-wires, etc., for each wing. With the relative proportions of the various parts of the wooden framing of the wings of the models as a basis, it was decided to make the main ribs of the large wings 1.5 inches in diameter for one-half their length, and have them taper from this size to one inch in diameter at the extreme point. After making allowance for the weight of these ribs, it was found that, if the cross-ribs were to be spaced no farther than ten inches apart, and the two end ones were to be made at least as wide as 1.5 inches in order to resist the end strain due to the stress of the cloth, the twenty-six intermediate cross-ribs could be only seven-sixteenths of an inch in diameter at the point where they crossed the main rib, and that they must be tapered to three-eighths of an inch in diameter at the front end and to one-fourth of an inch in diameter at the rear tip.

A trial wing, whose total weight was 30 pounds 2 ounces, was made up with the various parts of its frame of the above dimensions. Even upon inspection it appeared to be too flimsy to withstand the sudden gusts of wind which were certain to be met in actual practice. In order, however, to get some definite data as a guide, the wing was inverted and guyed in the same way that it was proposed to guy it on the aerodrome, and a uniform thickness of sand was then sprinkled over it to such a depth as to give it a load of 0.7 pounds per square foot. Even before one-quarter of the sand was sprinkled over it, it was seen that the wing was rapidly going out of shape, and it was feared that the full amount of sand would not only seriously distort it, but would even break it. The full quantity of sand, however, did not break it, but distorted it to such an extent that, had the pressure been due to its being propelled through the air, its serious change in form would have rendered it worse than useless. [p196] While the main ribs had shown a certain amount of deflection under the sanding test, the more serious distortion had been in the cross-ribs, the small guy-wires, which had been fastened to each cross-rib, becoming loose instead of tight, as had been expected, since the rib tended to increase its curvature instead of straightening out. This increase in the curvature of the cross-ribs was partly overcome by tying the guy-wire flat against the cross-rib for a distance of about 2 feet from the rear tip. But while this caused the guy-wire to tighten the general contour of the wing showed very little improvement, as the ribs now assumed a curve more or less like the letter S, the rear tip now being bent downward to form the tail of the elongated S.

From this sanding test it was seen that the cross-ribs must be materially stiffer, and a new set was accordingly made one-sixteenth of an inch larger in diameter at the various points of measurement. Upon giving the wing, equipped with these larger ribs, a sanding test it was found that, while there had been some improvement, it was entirely too flimsy, even when it been double-guyed by running a second wire on each cross-rib from the middle of the portion in front of the mid-rib to the middle of the portion behind the mid-rib. As the weight of the wing with these larger solid cross-ribs had now increased to more than 33 pounds, and the wing had proved itself altogether too weak for use on the aerodrome, it was evident that some other plan of constructing the ribs which would give greater strength for the same weight must be found. At first sight it might appear that the obvious way of increasing the stiffness of the cross-ribs was to employ a cross-section other than a round one, since material added to the depth of the rib is very much more effective than if added to the width. It must, however, be remembered that these cross-ribs were 11 feet long, and that, as the main mid-rib was 6 feet in front of the rear tips of the cross-ribs, with no intermediate bracing, except the light threads by which the cloth cover was attached, it was inevitable that, should the depth be made materially greater than the width, the rib would buckle sideways. Test ribs of I-beam form, which are later described, were constructed, but, although they proved exceedingly stiff, had to be discarded.

In view of these facts the obvious remedy appeared to be to make the rib hollow, and one cross-rib, 3/4 of an inch in diameter at the point where it crossed the main rib, tapering to 5/8 of an inch at the front and 3/8 of an inch at the rear tip, was accordingly constructed. Tests showed that this form of rib, which was about 10 grammes lighter than the 1/2-inch solid ribs, was much stiffer than anything yet constructed. But when a wing, with cross-ribs of this size placed 20 inches apart, was sanded it was found that, although a great advance in construction had been made, still further improvement was necessary before a suitable wing for the large aerodrome could be procured. [p197]

Before proceeding with the construction of any more complete wings, an extended series of experiments was made in order to secure ribs of proper lightness and strength. Various forms of metal tubes were tested; but, although aluminum seemed at one time to promise good results, it was found that hollow ribs could be constructed of spruce which were much stronger than aluminum tubes of the same weight. In order to determine more accurately what mode of construction would give the greatest stiffness and strength for a minimum weight, it was decided to make up some test pieces of different forms before making up complete ribs. For convenience in construction, these test pieces were made straight and shorter than the large cross-ribs. Each piece was tested by fastening it in the testing clamp with 1 metre of its length projecting horizontally, and attaching at its end a weight of 1 kilogramme. The deflection from the horizontal gave an index of the stiffness of the piece under examination.

The first test piece was a hollow square, 17 mm. length of side on the exterior, and 11 mm. length of side on the interior, the walls thus being 3 mm. thick. This weighed 73 grammes per metre and had small internal stiffening pieces, like the partitions in bamboo, glued into it 4 inches apart. A weight of 1 kilogramme at the distance of 1 metre gave a deflection of 56 mm. The second test piece was a duplicate of the first one, except that it had no internal stiffening pieces, and the weight per metre was made the same, 73 grammes, as formerly, by leaving the walls a fraction thicker. The deflection in this case was, as would be expected, exactly the same as in the first one. The first test piece, however, was superior to the second one in that it was stiffer against being crushed in by accident. The third test piece was a hollow cylinder, 22 mm. outside diameter and 17 mm. inside diameter, the walls thus being 2.5 mm. thick. The weight per metre was 91 grammes, and the deflection was 46 mm. The fourth test was made by taking two of the original solid cross-ribs, 12 mm. in diameter, and fastening them in the clamp side by side, with a length of 1 metre projecting. The weight per metre for the two ribs was 105 grammes, and the deflection produced on the two by 1 kilogramme at 1 metre distance was 115 mm. The fifth test piece was an I-beam of spruce, having a depth of 25 mm., with the flanges 12.5 mm. wide and the web 3 mm. thick. The weight per metre was 65 grammes, and the deflection was 26 mm. All of these test pieces were made of carefully selected straight-grained spruce.

It is readily seen that the test piece having the I-beam section weighed less than the hollow square in the first and second tests, and had a deflection of less than half. This I-beam section, however, did not show up so well when a longer piece was tested, for as soon as the length was made appreciably greater than a metre it began to twist, the twisting becoming more and more serious the [p198] greater the length, until with a piece 11 feet long, the full length of a cross-rib, the twisting was so serious as to make the rib practically useless. It was at first thought that this twisting might be overcome by making the webs slightly wider, and it would to a certain extent, but in looking ahead and planning how the cross-ribs were to be fastened to the main ribs, the I-beam section was seen to present so many difficulties that it was thought hardly worth while to spend time on further experiments with it. This decision was made all the more imperative by foreseeing the difficulty of bending the I-beam section to the curve which the cross-ribs were to have. In fact it had been found by experience that while many different forms of ribs could be bent to the proper curve by steaming and clamping them over a form and then drying them out while still clamped to the form, yet the grain of the wood varied so in different ribs, that of a dozen steamed and bent over the same form it was seldom that as many as three would have approximately the same amount of curvature when removed from the form after drying. If, however, the curve was formed in the ribs by making them in two parts, which were glued together and clamped up on the form while the glue dried, practically any number could be made which would have the same curvature when thinned down to the proper thickness of wall.

It was recognized at all times that the gluing together of the ribs not only entailed extra work, but introduced an element of uncertainty unless some kind of a varnish for the ribs could be found which would prevent any possibility of the glue becoming soft from moisture in the atmosphere or from the wings actually coming down into the water when the aerodrome was tried in flight. A search was therefore made for a varnish that was water-proof. A large number of different varnishes were tried, and one was finally found which, after repeated tests, seemed to be thoroughly good. Several test ribs were given three coats of this varnish, and were then kept immersed in water for 24 hours without the glue showing any signs of softening. It was therefore decided to follow the plan of gluing the ribs together and protecting them with three coats of this varnish, which seemed to possess the remarkable properties of being not only impervious to water, but also unaffected by the application of concentrated ammonia or of gasoline, either of which produces immediate softening when applied to ordinary varnishes.

Following the indication of these tests that the hollow, round section, 22 mm. outside diameter by 17 mm. inside diameter, would probably give the best cross-rib for the weight that it seemed possible to allow, a set of cross-ribs of this form was constructed and put in place in the large experimental wing, in which the former solid ribs had been tested. The wing was inverted and fastened into two posts at the angle it would have in flight, the guy-wires from the [p199] lower guy-posts of the aerodrome being represented by wires stretched from the posts. In actual use on the aerodrome it was proposed to have three main guy-wires running from each of the main cross-ribs to the lower guy-post, but in the test, which is now to be described, the wires which would have come nearest the body of the machine were left off to see what effect their removal would have on the wing.

The weight and dimensions of the wing, as set up, were as follows: Length of the main ribs, 24 feet; length of the cloth covering, 22 feet; width of the cloth covering, 11 feet; total weight, 29 pounds. The two main ribs (front rib and mid-rib) were solid, 3.5 cm. in diameter at butt, 2.5 cm. in diameter at tip, and tapering from the middle to the tip. There were twelve regular cross-ribs set 50.8 cm. (20 inches) apart, each rib being as above described, 11 feet long, 22 mm. outside diameter by 17 mm. inside diameter at the butt, and tapering from where they were attached to the mid-rib to the tip, and each weighing 300 grammes. There were two extra cross-ribs, one at the inner end next to the body of the machine and the other at the outer end. These were solid strips of wood 3.8 cm. wide by 1.2 cm. thick, made extra wide and stiff in order to withstand the strain of stretching the cloth covering. There was also a thin, flat strip at the rear edge, which connected together all the tips of the cross-ribs, holding them a uniform distance apart, and also serving to fasten the cloth. The main mid-rib was stiffened in a vertical direction by a system of small guy-wires drawn over short guy-posts about 6 inches high. With the wing inverted and fastened in the way above described, a weight of 2 kilogrammes placed at the inner rear corner produced a deflection of 26.7 cm. When the inner rear corner was pulled up by a spring balance until the balance registered 2 kilogrammes, there was an upward deflection of 41.3 cm. When the main mid-rib was held at the inner end, the pull of 2 kilogrammes, applied to the inner corner as before, caused an upward deflection of 25.4 cm. instead of 41.3 cm. This wing was afterwards given a sanding test under a weight of 0.7 pound per square foot. With fine guy-wires fastened from the front of the cross-ribs to the tip and drawn just taut, the ribs showed an average deflection of 9 inches at the tip under the above weight. When a small wooden guy-post was added under each of these small guy-wires, the same weight produced an average deflection of 5 inches at the tip of each rib under the same load. In a previous test of the wing, using hollow cross-ribs 16 mm. outside diameter by 10 mm. inside diameter at the butt, and only half as far apart as the later ones, a load of 1 pound per square foot on the wing produced an average deflection of 9 inches at the tip of each rib when the cross guy-wires on each rib were held up by short guy-posts, but when these short guy-posts were removed, the same load produced a deflection of nearly 25 inches at the tip of each rib. [p200]

Although this wing was a great improvement in every way over any of the previous constructions, it was felt that it was too weak for the large aerodrome. Further experiments were therefore made in order to secure a form of cross-rib which would meet the rigorous requirements imposed. An inordinate amount of time was spent in the construction and tests of various forms of rib, but as a result a satisfactory cross-rib was at last constructed of the form shown in Plate 66, Figs. 4–8, the dimensions at the three principal points, viz., first, where the cross-ribs join the front rib; second, where they cross the mid-rib; and third, at the rear tip, being given both for the intermediate cross-ribs and the end cross-ribs.

Following the plan employed by Nature in the construction of the bamboo pole, small partitions, approximately one millimetre thick, were placed every three inches in the thin, hollow rib to keep it from being crushed. The partitions were glued in place when the hollow rib was glued together on the form around which it was bent and clamped until the glue dried. Longer blocks were also inserted in each of the intermediate ribs at the point where it crossed the main rib and also at the front end where it was attached to the front rib. In the end ribs blocks were also inserted at the points where the cross-braces were fastened to them for resisting the end stress due to the cloth covering.

Upon making up one of these ribs and testing it, it was found to possess remarkable stiffness, so much so that it was thought probable that it was as stiff in proportion to its size as the best thing that Nature had produced in the bird’s wing. A large quill from the wing of a harpy eagle was therefore stripped and the large end clamped in a special holder, and measurements were made of the deflection produced by weights at various distances from the clamp. As the main mid-rib of the wing of the aerodrome is placed approximately at the point of the center of pressure, the bending action on the cross-ribs may be assumed to act on a lever arm from the mid-rib towards the front, and from the mid-rib towards the rear in the cases of the pressure on the front and rear portions of the wing, respectively. In testing these cross-ribs, therefore, against the quill, the rib was clamped at the point where it crosses the mid-rib of the wing, and measurements were made of the deflection produced by weights placed at various distances from the point of clamping both front and rear.

The quill on which the measurements were made was 19.5 inches long and had a gradual curve, the highest point of the curve being about the center of the length of the quill, and the depth of curvature being about 2 inches. When the butt of this quill was placed in the clamp the tip stood 17 cm. above the horizontal. The hollow spruce rib, when clamped at a point 5 feet from the tip (the point from which it tapers in both directions) had its tip 2.2 cm. above the horizontal, there being very little curve in that portion of the rib. The quill weighed 4 grammes when stripped and 18 inches of it projected from the [p201] clamp which held it during the tests. The rear portion of the spruce rib projected 5 feet from the clamp, being thus 3.3 times as long as the quill, and it weighed 120 grammes, the weight for the larger size having, therefore, increased slightly less than the cube of the length.

The results of the tests of both the quill and the rib are given in the following table. The approximate cross-section of the quill at the point of clamping, the middle and the tip are shown in diagrams ‹A›, ‹B› and ‹C›, respectively, of Plate 67. The cross-sections of the rib at the corresponding points are shown in diagrams ‹D›, ‹E› and ‹F›. The cross-sections of the quill, enlarged five times, are shown in diagrams ‹A′›, ‹B′›, and ‹C′›.

QUILL FROM THE REMIGES OF HARPY EAGLE.

Weight, 4 grammes; length, 45 cm.; tip, 17 cm. above butt when the latter is horizontal.

Point of Weight application Absolute in terms Deflection of weight weight in of in terms in terms of grammes. greatest of length. length. weight. 0.39 1050 1.0 .38 0.445 605 0.58 .38 0.56 405 0.39 .38 0.75 210 0.20 .38 0.95 77 0.075 .38

HOLLOW SPRUCE RIB.

Weight, 120 grammes; length 153 cm.; section, rectangular; tip, 17 cm. above butt.

Point of Weight application Absolute in terms Deflection of weight weight in of in terms in terms of grammes. greatest of length. length. weight. 0.39 15,000 1.0 .11 0.445 11,400 0.76 .11 0.56 7,900 0.53 .11 0.75 4,000 0.27 .11 0.95 2,000 0.135 .11

In each case the unit of length was the portion extending beyond the clamp; the unit of weight, the greatest weight employed to produce the deflection. It should be noted, however, that the relative deflection was quite different in the two comparisons. In the case of the quill the deflection was 17 cm. in 45 cm., or 38 per cent; in the case of the rib it was 17 cm. in 153 cm., or 11 per cent. In the case of the rib at the point 0.39 the absolute weight was 15,000 grammes, the relative weight unity and the deflection in terms of length 0.11. While no rigorous comparison can be instituted, since the rib was not deflected nearly as much proportionately as the quill, yet the general inference is that while the rib was not intended to be, and was not as elastic proportionately as the quill, it was probably at least as strong in proportion to its weight. Briefly summarizing these results it will be noted that the spruce rib was about 3.3 times the length and 30 times as heavy, while it was 15 times as stiff near the butt and 26 times as stiff at the tip, as the quill.

As this test on the rib for the large wings had apparently shown that the plan of constructing the ribs in the form of a hollow square secured maximum strength for minimum weight, it was decided to construct a few sample ribs after the same plan for the wings of the new quarter-size model of the large aerodrome, and to test, these ribs in a similar manner. The following table shows the results of the test on one of these ribs:

Total length of rib = 80 cm. Curve = 1 in 18. Highest point of curvature = 0.25 from front. Section of rib = 10 mm × 14 mm. at the point of attachment to mid-rib, tapering to 8 mm. × 12 mm. at the front point and to 7 mm. × 2 mm. at the tip. The rib was clamped with the tip projecting [p202] 46 cm. and was weighted at different percentages of its length to such an extent that it was deflected 11 per cent of its length, or 5 cm. The weight of the 46 cm. length of rib which projected from the clamp was 11 grammes, the whole rib weighing 22 grammes and balancing on a knife edge placed at the point where it was clamped.

Point of Absolute Weight in Deflection application weight terms of in terms of weight in greatest of in terms of grammes. weight. length. length.

0.39 7680 1.0 0.11 0.445 5980 0.78 0.11 0.56 3680 0.48 0.11 0.75 2300 0.30 0.11 0.95 1100 0.143 0.11 1.00 930 0.121 0.11

A lighter rib than the above, which was constructed at the same time, was also tested with the results shown in the following table. This rib was also 80 cm. long, but was only one-half the linear dimensions in section of the rib previously tested. The rear portion of it projected 46 cm. from the clamp. The total weight of the rib was 11 grammes, or 5.5 grammes for the 46 cm. on which the measurements were made.

Point of Absolute Weight in Deflection application weight in terms of in terms of weight grammes. greatest of in terms of weight. length. length.

0.39 1400 1.0 0.11 0.445 1100 0.785 0.11 0.56 700 0.50 0.11 0.75 400 0.275 0.11 0.95 250 0.178 0.11 1.00 220 0.157 0.11

A still lighter rib of the same length, weighing 9 grammes, suitable for use in the wings of the quarter-size model, was constructed and a set of tests was made on it with the following results. As in the above test, 46 cm. of the rear portion of it projected from the clamp which held it.

Point of Absolute Weight in Deflection application weight in terms of in terms of weight grammes. greatest of in terms of weight. length. length.

0.39 1450 1.0 0.11 0.445 1150 0.795 0.11 0.56 740 0.51 0.11 0.75 380 0.262 0.11 0.95 210 0.145 0.11 1.00 180 0.124 0.11

Among quite a number of different forms of cross-ribs which were constructed of a size suitable for use in the model aerodrome, but made primarily for use in tests to determine the best form to employ, may be mentioned the following, in which both ribs were seven-sixteenths of an inch outside diameter and five-sixteenths of an inch inside diameter. One was filled with elder pith, formed up into a round rod that just fit the interior of the hollow rib, and was glued into it when the rib was glued up. The other rib was left hollow. Upon testing these by suspending weights at different points, the rib without [p203] the pith showed a slightly less deflection than the one with it, it happening probably that the wood in one case was a little stiffer than in the other, although they were carefully selected to be as nearly alike as possible. The rib with the pith in it weighed 34 grammes and the one without it weighed 30 grammes. It was inferred from this test that the placing of a light pithy material in the interior of the ribs would have no good effect, and would not only add weight, but also complicate the construction. The reason for making this test with pith in one of the ribs was that it was thought probable that the rib flattened out somewhat when it was deflected under a load, and that the pith stiffened with the glue with which it was fastened in, might lessen this.

As the cross-rib described above, which was tested on October 23, 1899, seemed in every way suited for use in the wings of the large aerodrome, a complete wing equipped with similar ribs but of slightly changed dimensions, as shown in Plate 66, Fig. 5, was immediately constructed. As previous tests had shown that the wing covering did not “flute” or “pocket” to any considerable extent even when the ribs were as much as thirty inches apart, only ten cross-ribs were used in this wing. The eight intermediate cross-ribs were of the form described above, but the ribs at either end of the wing were made of a larger cross-section and otherwise stiffened in order to resist the strain of the tightly stretched cloth covering.

On April 13, 1900, a final sanding test was made on this wing, guyed in a manner similar to that used in the aerodrome, in which the following results were obtained:

SANDING TEST OF LARGE WING.

Area, 260 sq. ft.; weight of wing, 29 pounds; weight of sand on wing, 231 pounds; total weight supported by wing, 260 pounds, or one pound per square foot.

Deflection of cross-rib, numbering from inner edge to extreme outer edge of wing--

Number of rib. Deflection. Inches. 1 (Heavy end rib) 5.5 2 9.5 3 11.75 4 12.25 5 12.5 6 12.75 7 12.9 8 13.0 9 12.0 10 (Heavy end rib) 9.75

The weight of sand put on the wing in this test was 1.5 times as great as the pressure which at this time it was expected would be imposed upon it in flight, and was in fact 1.2 times as great as the normal pressure when supporting the aerodrome as finally constructed. Even under this weight the greatest deflection noted in terms of the total length of the rib was less than 0.10, showing that the elastic limit of the rib was far from being reached. [p204]

As this test seemed to indicate that the wings constructed in this manner were certainly strong and rigid enough for use on the aerodrome, and that immediate further improvement could hardly be made, three similar wings were at once constructed to complete the set. Somewhat later two additional wings were provided, so that when the large aerodrome was taken to Widewater on the Potomac in 1903 one and a half complete sets of wings were on hand, which seemed to be ample to provide for any emergencies that might arise.

Each of these wings had, as is clearly shown in the drawings, Plates 53 and 54, two main ribs, which formed the main strength of the framework and gave the wing longitudinal rigidity. To the main front rib were attached the cross-ribs and the pieces for the curved extension later described. The mid-rib extended across the cross-ribs, parallel to and about 5 feet behind the front rib, this being approximately the line in which lay the center of pressure of the wing. It was upon this rib, therefore, that the greatest strain would fall.

The mid-rib, Plate 66, Fig. 2, was 731.5 cm. (24 ft.) long, having at the butt an outer diameter of 38 mm. (1.5 in.) and an inner diameter of 25 mm. (1 in.), the walls being, therefore, approximately 6.5 mm. (0.25 in.) thick. From the butt to the middle point the section was uniform, but from this point it had a taper of one-twenty-fourth of an inch to the foot, so that at the tip it had an outer diameter of 25 mm. (1 in.), the thickness of the wall being unchanged. At the butt end a wooden block 8 inches long was glued inside the rib, and at uniform distances of 75 mm. (30 in.) 10 smaller blocks were glued in where the cross-ribs were attached. The main front rib was of the same form and size, except that it was some 2 inches shorter and had no blocks, except the long one at the butt, glued in it.

To these main ribs were attached, in the manner later described, the 10 cross-ribs, to which the cloth cover was attached. The 8 intermediate cross-ribs have already been described in connection with the tests. The cross-ribs at the end of the wings, upon which greater lateral strains would come from the stretching of the cloth, were made of the larger cross-section shown in Fig. 8 of Plate 66. Additional longitudinal stiffness was provided by gluing a strip 2 mm. thick between the upper and lower halves, as shown in the section. These end ribs, as well as those next to the ends, had small blocks glued into them where they were crossed by the diagonal braces, in addition to the small partitions 1 mm. thick, which were glued into the ribs every 3 inches to prevent crushing, and the blocks 2.5 and 3 inches long respectively, where they were attached to the front rib and to the mid-rib. At the extreme rear edge of the wing the cross-ribs were attached to the small “D”-rib, which served to hold the ribs at equal distances and to keep the cloth cover stretched tight. This “D”-rib, as shown in Plate 66, Fig. 3, had semi-circular walls 4 mm. thick, 21 mm. in diameter, to the edge of which was glued a flat strip 3 mm. thick. [p205]

As originally designed the wings had a curve of only 1 in 18, the main front rib forming the leading edge of the wing. Later, however, it seemed desirable to “quicken” the curve and at the same time give the wing a sharper leading edge. This was accomplished by attaching to the front rib, at the points where the cross-ribs joined it, properly curved wooden pieces of the form shown in Plate 66, Fig. 10, over which the cloth cover of the wing was stretched. The curve of the wing after the addition of this extension is shown in Plate 66, Fig. 4, and was a curve having a rise of approximately 1 in 12, with the highest point .25 from the front end.

On account of the large size of these wings and the consequent difficulty in handling them it was necessary to construct them in such a manner that they could be easily taken apart, rolled up, transported to the house-boat or any other point where they might need to be used, and then quickly reassembled. After much experiment as to the best means of constructing them, the following plan was devised. The cloth covering was permanently fastened to the front rib, to which were attached the front extension pieces by means of small metal clips secured by small wood screws. On the rear edge of the front main rib, at a uniform distance of 30 inches apart, 10 small metal horns of 1-mm. tubing, 5 cm. long, each brazed to an independent clamping thimble, as shown in Fig. 9 of Plate 66, were fastened. The front end of each of the cross-ribs was slightly rounded out to fit the front main rib, and in the wooden block which was glued in this end of the cross-rib a hole was bored to fit these horns. Each of the cross-ribs was then pushed over its proper horn and against the front main rib, and the cloth covering then drawn back toward the rear tips of the cross-ribs. In the extreme rear edge of the cloth covering a seam was made, and in this was inserted the “D”-rib already described. The cloth was then tightly stretched and a wood screw forced through the “D”-rib and into and through the metal ferrule at the tip of the cross-rib. Near the inner and outer edges of the cloth covering eyelets were placed about 6 inches apart, through which small cords were then inserted and tied to the end cross-ribs. The main or mid-rib was then placed on top of the cross-ribs and fastened to them with wood screws, and the cross-braces were then fastened on the top of the wing, as shown in Plate 54. The frame of the wing was stiffened horizontally by cross guy-wires which passed from each cross-rib, at the point where the mid-rib crossed it, to the adjoining cross-rib, at the point where it was connected to the front rib. Each of the main ribs was individually guyed, in the manner clearly shown in Plate 52, in order to stiffen it in the vertical direction, the fittings for these guy-wires being shown in detail in Figs. 11–15 of Plate 66. Finally, small guy-wires were run from the front end of the cross-ribs over a guy-post 12 inches high at the point where the cross-rib crossed the mid-rib to the rear tip of the cross-rib. These cross guy-wires were regulated in [p206] tightness by raising and lowering a screw in the slot of the head of which they rested, and which was threaded in the end of the small guy-post. Upper and lower guy-wires, running from the main ribs to the guy-posts on the aerodrome, as already described, and as is clearly shown in the drawings, Plates 52 and 54, completed the guy-wire system for the wings, except for the “drift wires,” which for the front wings were run from the lower side of the mid-rib to the bowsprit at the front of the machine, and for the rear wings to the main frame.

Each wing when completely assembled weighed approximately 29 pounds, and had a rectangular surface 22.5 by 11.5 feet (measured on the chord of the curve), or 260 square feet, making the weight per square foot equal about 50 grammes, rather less than 1.5 times as much per square foot as the wings for the steam-driven models. The total supporting surface of the aerodrome was 1040 square feet, and as the aerodrome when equipped for flight weighed, including the aviator, 850 pounds this gave 1.22 square feet to the pound, or 0.82 pound to the square foot. Although this was a somewhat larger proportion of weight to supporting surface than it had originally been expected to have, there is every reason to believe that it was sufficient, for the quarter-size model, when weighted so that it had 1.22 square feet to the pound, flew well, as will later appear.

[p207]