the front axle on a single transverse spring, thus reducing the number of springs on the car to three.

When the springs have been depressed by an obstacle in the roadway, they only return to the position of equilibrium after a number of oscillations of decreasing amplitude have taken place. It is advisable to spare the vehicle this continued oscillation, as at high speeds it causes the wheels to leave the ground, and consequently reduces the effective power of the motor. M. Truffault has taken out a patent for an arrangement to remedy this defect. The friction between two metal surfaces prevents the oscillations from arising. He tried a spring fitted with this damping action on a quadricycle, which carried his son to victory over the kilometre at Deauville in 1901. This spring has given very good results, enabling one to travel rapidly even over the worst of paved roads.

Effect of the Nature of the Tyres.

The experiments of M. Michelin have shown that the tractional resistance is reduced from 15 per cent. to 30 per cent., according to the nature of the road, by the use of pneumatic tyres in place of metal tyres. He explains this by the well-known saying, 'Le pneu boit l'obstacle.' Baron de Mauni in a recent work has given an account of some experiments which he made on different tyres, particularly pneumatics. He showed that if two wheels with tyres of equal widths supported equal loads, the one that has the greater arc of contact with the ground will travel better than the other. With a rigid tyre such extended contact can only be secured by increasing the diameter of the wheel, which is impossible beyond certain limits, so that the tyre will sink into the ground by an amount proportional to the weight carried. With rubber tyres the increased area of contact is due to the elasticity of the material and not to the increased diameter, so that the wheel does not sink into the road.

Professor Baker's experiments seem to show that on good roads the width of tyre has little effect on the resistance, and that even on bad roads the advantage lies sometimes with the wide tyres and sometimes with the narrow ones, according to circumstances. Arguments have been advanced in favour of both wide and narrow tyres, but nothing very definite seems to be known on the subject; according to M. Michelin, if we reduce the width of the tyre we reduce the adhesion to the ground, which is already little enough. As a case in point, he mentions that M. Serpollet, in order to attain a speed of 120 kilometres (75 miles) per hour on the Promenade des Anglais in 1902, had to deflate his tyres, and thus get a larger surface of contact with the ground. Besides, in order to get a narrow tread it is necessary to give to the tyre a form other than circular, and this shape can only be retained at the expense of its flexibility. Consequently a tyre of this description will be subjected to greater internal friction in its fabric than one naturally circular in section, and the energy wasted will be therefore greater.

The whole question, however, is very much open to discussion, and the present Congress may offer to the opposing schools an opportunity of coming to some understanding.

Resistance of the Air.- Study of Forms to diminish this Resistance.

The air resistance is a retarding force of the highest importance, especially where speed is concerned, and there is unfortunately great

uncertainty both as to the formulæ to be applied, and the values of the coefficients which appear in them.

The best known formula is

R = KSV2

in which R equals resistance in kilogrammes.

S equals projected area in square metres of total surface of vehicle on a plane normal to direction of motion.

V equals velocity in metres per second.

Ka numerical co-efficient which varies between very wide limits according to the form and the speed of the vehicle.

The formula by M. Desdouits, R = KV, is sometimes preferred, as it is more correct for high speeds.

The different values given to K in the first formula may be due to the varied conditions under which the experiments were made.

Signor Canovetti had made some experiments at Zossen to determine the value of K. He had a copper wire, 380 metres long, stretched between the summit of the fortifications at Brescia and a point in the plain, about 70 metres below. Along this wire different surfaces were allowed to descend freely. A circle, with a surface of '073 square metres, moving with a velocity of 12 metres per second, gave a resistance of 84 grammes. The same circle, having a spherical cap in front, offered a resistance of only 21 grammes. When this hemisphere was followed by a cone, whose height was five times its diameter, the resistance fell to 13 grammes, or one-sixth of that of the plane circle. With this same solid, turned the other way about-that is, with the apex of the cone towards the direction of motion-the resistance rose to 18 grammes.

Signor Canovetti has recognised that a rectangular surface, placed with its long sides horizontal, offers a sensibly greater resistance to the air than when its short sides are horizontal. His experiments seem to show that the coefficient K diminishes somewhat as the speed increases, but investigations carried out at Zossen point to the conclusion that the resistance may increase tenfold when the velocity is only tripled. It is thus clear that air resistance is a matter of no small importance when speeds up to 60 or 80 miles an hour are attained. At 85 kilometres (53 miles) per hour, the energy required to overcome the air resistance on a vehicle, with an opposing surface of 1 square metre (1,550 square inches), may be 7, 11, or 20 horse-power, according to the coefficient K given as 0.0288, 0.0648, or 0.116 by MM. Forestier, Bourlet, or Thibault.

The question then arises, What is the best shape for a car? The answer depends on several things-as, for example, the necessity of placing the radiator in such a position that it may be efficiently cooled by the air rushing through it. Only general principles may be laid down. The front of the car ought to taper, and the back be more pointed still, like the form of a fish: transverse rectangular surfaces that cannot be dispensed with, should, as far as possible, have the longer sides vertical; and it is well to have doors on the car to prevent the air from rushing in between the dashboard and the seat.

These conditions are quite neglected in most of the present-day cars. Particularly is this the case in the Coffin Head,' that unlovely affair so much in vogue-a flat surface directly opposed to the air pressure. With a radiator of the honeycomb type, a transverse position is necessary for

cooling purposes, and Signor Canovetti has shown evidence that a perforated surface will offer less resistance to the air than a plane one of similar area. This is not of much advantage with an automobile as the air, after having passed through the holes in the radiator, meets with further obstacles in the mechanism inside the hood.

With regard to the working parts situated under the car, these should be made by the aid of inclined planes to cut the air rather than oppose it. M. Lavergne commends the suggestion of M. Forestier that differentshaped bodies should be mounted on an electric chassis, and the total resistance of chassis and body accurately measured, so that a really practical model could be designed.

Poner required by Automobiles.

Under this heading M. Lavergne has shown the enormous reduction of weight per horse-power that has taken place during the last eight years. In 1895, Levassor made the run from Paris to Bordeaux in a 4 horse-power car weighing about 1 ton, or 1 horse-power per 550 lb. dead weight. In 1896 this weight was reduced to 365 lb. per horse-power; in 1900 it fell to 90 lb. per horse-power. In the recent Paris-Madrid race M. Gobron Brillie appeared with a 100 horse-power car, the weight of which represented only 22 lb. per horse-power. This weight has been still further reduced in the case of motor bicycles, reaching as low a figure as 17.5 lb. per horse-power.

But there is not a corresponding increase in speed. In 1901 M. Fournier made over 53 miles per hour with a 28 horse-power Mors; last year M. de Knyff only slightly exceeded 58 miles per hour with a 70 horse-power motor; that is, an additional 40 horse-power.

To what must this relatively small increase of speed be attributed ? Air resistance is responsible for some increase but certainly not all.

An extremely powerful motor must be accompanied by a comparatively heavy load, otherwise the wheels do not 'bite' well and energy is wasted. It is well known that the modern racing-car skims along the surface of the course, without sufficiently close contact between the wheels and the ground; in any case driving wheels should be more heavily weighted and springs made less elastic. To reduce the power lost in vibration, the engine should be more perfectly balanced, and, if necessary, the fly-wheel and motor itself made heavier. 'Who shall say,' M. Lavergne concludes, 'whether, instead of building very powerful yet extremely light motors— the durability of which is questionable-it would not be better to rest content with a vehicle of smaller power, and use it more effectively?'

IV. Negotiations with the War Office.

At a Committee Meeting held at the Society of Arts on May 15, 1903, it was proposed that as the expenses of this research were extremely heavy it would be advisable to approach the Mechanical Transport Committee of the War Office, in order to see if they would conduct the experiments with heavy traction, as they had at their command various powerful motors and traction engines, together with the necessary variety of wheels. The Transport Committee in return would have the use of the British Association recording instruments for their own experiments. This British Association Committee would have access to the information obtained which was of a scientific character with a view to

publication, but it would not concern itself with data relating to the actual waggons and other matters of a purely military character.

As a result the Transport Committee replied favourably, and arrangements, it is hoped, will now be made by which important work will be carried on by that Committee, thereby avoiding the very heavy expense to meet which it is difficult to raise funds from private sources.

Small Screw Gauge.-Report of the Committee, consisting of Sir W. H. PREECE (Chairman), W. A. PRICE (Secretary), Lord KELVIN, Sir F. J. BRAMWELL, Sir H. TRUEMAN WOOD, Major-Gen. WEBBER, Col. WATKIN, Lieut.-Col. CROMPTON, A. STROH, A. LE NEVE FOSTER, C. J. HEWITT, G. K. B. ELPHINSTONE, E. RIGG, C. V. BOYS, J. MARSHALL GORHAM, O. P. CLEMENTS, W. TAYLOR, Dr. R. T. GLAZEBROOK, and MARK BARR, appointed to consider means by which Practical Effect can be given to the introduction of the Screw Gauge proposed by the Association in 1884.

THIS Committee was originally appointed at the York meeting of the British Association in 1881, and, after considerable labour, they made their final report at the Montreal meeting in 1884, recommending a very useful series of small screws, which were very generally adopted for watch and electrical apparatus. At the Ipswich meeting of the British Association in 1895 the Committee was reappointed to consider complaints that screws of the British Association thread proposed by the Committee in 1884 were not interchangeable. It appeared to the Committee that the difficulty arose from want of some ready means of constructing gauges for testing the screw thread, and they endeavoured, during the years 1896-9, to remedy this by the construction of a series of such gauges. The edges of the thread, as is well known, are rounded at the crests and roots, and great difficulty was experienced in obtaining satisfactory gauges for such a form, while it was stated that a flat-topped thread could be accurately made, and gauged with comparative ease. At the Dover meeting, in 1899, this Committee reported recommending that they should be reappointed for the purpose of considering whether the British Association form of thread for small screws should be modified. This recommendation was adopted, and as a result the Committee reported at the Bradford meeting, in 1900, that it was desirable to replace the present form of screws from No. 0 to No. 11, by one having a flat top and bottom to the thread. It was pointed out that in the belief of the Committee such screws would, owing to the inevitable rounding at the edges, be interchangeable with the old stock in the majority of cases, and that only where great care had been taken to work closely to the old standard would any difference be noticed, so that practically while the B.A. small screw gauge had a flat-topped thread, the B.A. screws would still have rounded tops and bottoms. After making recommendations to the above effect the Committee was reappointed to obtain a set of the proposed screws, with tools and gauges for a comparison with the present ones. Additional members were added, and new light was thrown on the matter by their assistance. On the one hand, it appears that gauges can be constructed readily and accurately only if the thread be flat topped. On the other, that screws made with any ordinary form of screwing tackle will have round tops, but that the form of these tops will vary and may vary to such an extent as to prevent the inter

changeability of the screws. The resolution of 1900, modifying the form of the thread, was intended to apply to the gauges only, and it was supposed that the roots and crests of the thread in screws formed by dies or in nuts formed by taps would still be rounded, and the form of the thread would thus approximate very closely to that of the Montreal definitions. However, in view of the facts which have been put before them, the Committee now think it best to state explicitly that they do not propose to alter the form of the B.A. screw thread, and that they desire to withdraw the recommendation accepted at Bradford in 1900.

The original form of the British Association screw thread was laid down in the Montreal Report, 1884, to the following effect, viz. :

The angle of the thread was defined to be 47.5°.

The values of the pitches and of the external diameters of the screws were scheduled in millimetres; the depth of the thread was defined to be sixteenths of the pitch, and the radius of the rounding was to be the same at root and crest. The radius of the rounding was very nearly two-elevenths of the pitch by this rule.

This definition, closely adhered to, led to inconveniently small fractions in those quantities which had to be calculated from the scheduled dimensions, but this may be avoided if we replace the definition by an equivalent schedule of dimensions of all parts from which the insignificant figures are omitted.

Retaining the angle of 47.5° in all cases, we adopt the following schedule: