T HE apparatus here described has been designed for the rapid collection in duplicate of relatively large volumes of air to be examined in the laboratory for such gases as carbon monoxide, hydrogen sulphide, oxides of nitrogen, etc. Minute amounts of these gases are of hygienic significance and the standard methods of analysis require the aspiration of large quantities of such air through appropriate absorbents. It is not practicable as a rule to make the analysis on the spot, particularly if the air is rapidly changing in composition, owing to the time required for aspiration and the special apparatus necessary to complete the analysis. A situation might arise, for example, in a sewer, in an engine-room or after gunfire in a turret aboard ship when it is imperative to take large samples speedily, so that a representative portion within a short-time interval may be obtained. The advantages of col^ lection in duplicate are obvious. The present apparatus is designed to fulfil such requirements. The apparatus is portable, protected against breakage under shipment, and air-tight. It consists of a wooden case 17 inches long, 12 inches deep and 141 inches high, enclosing two glass bottles each of 2,000 cc. capacity, with gas-tight glass stopcocks, and fitted with a standard thermometer. The bottles are connected by glass and metal tubes with theinlet of a brass suction pump of special design. The writer has determined on the basis of a large number of tests that each stroke of the suction pump draws about 800 cc. of air through the two bottles at the rate of about 1 stroke per second. This was observed by connecting the two lower stopcocks to a gas holder and exhausting a definite quantity of air from the latter through the bottles with the pump. The course of the air in sampling is as follows: Through stopcocks e 1 to bottles, through stopcocks e 2 to the reducing T-tube g to h and to pump. The bottlers rest on felt pads and are held firmly in place by copper straps encased in felt. The rubber stoppers are carefully wired in place. The pump is readily accessible for interior inspection. There is an outside carrying case provided for shipment not shown in 'the sketch. This has a protective arrangement of springs at the bottom and also on the under side of the cover; the sides, cover and bottom are covered with very heavy felt padding. The inside case is therefore guarded against sudden jar in handling and no further packing material is needed. Thorough tests for tightness were made. There may of course be marked changes in pressure of the contained sample as a result of temperature variations; thus, a drop from 70 F. to 32 F. would entail a fall