When speed rises, however, and the throttle is sufficiently opened, the pressure within the apparatus falls and affects a spring-pressed diaphragm K, which actuates a piston valve controlling the air passages L, so that this valve opens to the atmosphere more and more with increasing pressure reduction, and additional air thus flows into the carburettor and mixes with the air and petrol entering through F. By this device the required proportion of air to petrol is maintained through a comparatively large volume range. This 'change of air admission is rendered necessary because of the difference between the laws of air and petrol flow. In order to give a sufficient weight of petrol at low speeds when the pressure drop is small, it is necessary to provide a somewhat large area of petrol jet. When suction increases owing to high speed, this large area discharges too much petrol, and so necessitates a device, such as that described, which admits more air. A still simpler device is adopted in many carburettors—that of an additional air inlet valve, kept closed until wanted by a spring. Fig. 4 shows a diagrammatic section as used in the Vauxhall carburettor. Here the petrol jet and primary and secondary air passages are lettered as before. The same effect is produced by devices which alter the area of the petrol jet or increase or diminish the number of petrol jets exposed as required. Although engine designers have succeeded in pro-portioning mixture through a considerable range of speed and charge demand, so as to obtain effective power explosions under all these conditions, yet much remains to be done to secure constancy of mixture at all speeds. Notwithstanding much which has been said as to varying mixture, there is only one mixture of air and petrol which gives the best results—that in which there is some excess of oxygen, more than sufficient to burn all the hydrogen and carbon present. 

It is necessary to secure this mixture under all conditions, not only to obtain economy in running but also to maintain purity of exhaust gases. Most engines at certain speeds discharge consider-able quantities of carbonic oxide into the atmosphere with their exhaust gases, and some discharge so much as to give rise to danger in a closed garage. Carbonic oxide is an extremely poisonous gas which should be reduced to the minimum in the interests of the health of our large cities. The enormous increase of motor traffic makes it important to render the exhaust gases as pure and innocuous as possible. Tests were made by the Royal Automobile Club some years ago which clearly showed that carbonic oxide should be kept down to 2 % and under when carburettors were properly adjusted. Subsequent experiments have been made by Watson, which clearly prove; that in some cases as much as 3o% not only discharge purer exhaust gases but would work on very of the whole heat of the petrol is lost in the exhaust gases by ins- much less petrol than they do at present. Practically all modern petrol engines are controlled by throttling the whole charge. In the earlier days several methods of control; were attempted: (r) missing impulses as in fig. i of the Daimler engines; (2) altering the timing of spark; (3) throttling petrol supply, and. (4) throttling the mixture of petrol and air. The last method has proved to be the best. By maintaining the proportion of explosive mixture, but diminishing_ the total volume admitted to the cylinder per stroke, graduated impulses are obtained without any, or but few, missed ignitions. The effect of the throttling is to reduce compression by diminishing total charge weight. 
..."