_Suction_ of the film against a perforated back plate has been found a very successful means of securing flatness. Suction at the moment of exposure may be produced by the action of a bellows, which has been compressed beforehand by the camera-driving mechanism. Continuous suction can be produced either by a continuously driven pump, or by a Venturi tube placed outside the fuselage. The Venturi tube (Fig. 55) consists of a pipe built up of two cones, placed vertex to vertex, to form a constriction. When air is forced through this at high velocity suction is produced in a small diameter tube taken off at the constriction. A suction of two centimeters of mercury, acting through holes about one centimeter equidistant from each other in the back plate, has been found adequate to hold flat a film 18 × 24 centimeters. One merit of suction applied only at the moment of exposure is that the film-driving mechanism does not have to work against the drag of the suction. Continuous suction, on the other hand, gives a longer opportunity for flattening out kinks in the celluloid, and easily permits movement of the film during the exposure, either for the purpose of permitting a longer exposure or for the purpose of preventing distortion due to the focal-plane shutter. A disadvantage of continuous suction is the production of minute scratches on the celluloid surface as it drags over the suction plate. These are ordinarily too small to cause trouble, but may show up when printing is done in an enlarging camera. [Illustration: FIG. 55.—Venturi tube on side of plane.] =Static discharges= are produced by the friction of the celluloid against the pressure back or other surfaces with which it comes into contact. They show in the developed film as branching tree-like streaks (Fig. 56) and in cold dry weather may be numerous enough to ruin a picture. The discharges are noticeably less frequent with film coated on the back with gelatine (“N.C.”), but the extra gelatine surface is extremely undesirable. When handled by developing machines, as large rolls must be, this back gelatine surface becomes scratched and bruised in a serious manner. Plain unbacked film is much to be preferred if the static can be obviated. To avoid static, it is necessary to provide for the immediate dissipation of all acquired electrical charges. Experiments made by the United States Air Service have shown that nothing is so good as rather rough cloth, thoroughly impregnated with graphite, held in close contact with the celluloid during as great a portion of its travel as possible. In the United States Air Service film camera which uses suction through a perforated back plate, the plate has been covered with thin graphited cloth, and similar cloth sheets are pressed against the film rolls by sheets of spring metal (Fig. 65). In cameras with this equipment no trouble has been experienced with static. [Illustration: