WATER— Organic Carbon and Nitrogen. 971

It will save some trouble in practice if this tableconsists of the logarithms of the numbers repre-senting the capacity of the tube in c.c.’s rather thanof the natural numbers. It is sufficient to take thelogarithms to five decimal places.

To use the apparatus, pour mercury into thereservoir, t, and allow it to pass the pinch-cock, r,until it has entirely filled the measuring tube anddrips from the capillary tube at g. The cock at f isthen closed, and the laboratory vessel being placedin position is filled with mercury, either by exhaust-ing the air through h , or better, by putting one endof a short piece of caoutchouc tube in it beforeplacing it in position, and then drawing air throughthe tube like an inverted siphon. The caoutchouccan then be removed, and any remaining bubblesdrawn out through k. Both being filled completelywith mercury, the faces of the steel caps are greasedwith a little resin cerate, softened if necessary bythe addition of a little oil; they are carefully placedin contact and firmly clamped together. On openingthe cocks, f and b, mercury should pass freely throughthe capillary communication, and on closing h andlowering t to its lowest position, r being still open,no leakage should take place into acd. All beingready, transfer the gas to be examined to the labo-ratory vessel, and with a small pipette having analmost capillary delivering tube, the end of whichis curved upwards, introduce two or three drops ofa saturated solution of bichromate of potash. If thesolution becomes green, sulphurous acid is present,and must be absorbed by the bichromate, of whichmore must be added if necessary. The action ofliquids on the gas may be much facilitated by gentlyshaking the mercury trough in which the vesselstands, or rather by making the stand on which thetrough is supported vibrate so as to throw the liquidup on to the sides of the laboratory vessel, and thusexpose a large wet surface. After about five minutesthe whole of the sulphurous acid will have beenabsorbed. During this time, and at any time duringwhich the gas is exposed to the action of a reagent,mercury is allowed to pass slowly into the laboratoryvessel by the capillary connection, to mix the gasand remove liquid from the mouth of the capillarytube. Having insured the absence of sulphurousacid, open both stop-cocks, yand h, lower the reser-voir, t, and thus drive the gas over into the measuringtube. Close the stop-cock, h, as soon as the liquidfollowing the gas is within about 10 m.m. of thestop-cock. Then by raising or lowering the reser-voir, t, bring the surface of the mercury in themeasuring tube to coincide, as seen in the telescope,with a convenient graduation, the capacity of theapparatus down to which is known. The flow ofmercury is controlled by the pineh-eook, r. Thisdone, observe the position of the mercury in thepressure tube and measuring tube, the temperatureof the water in the jacket, and the barometric pres-sure of the air; reading pressures to a tenth of amillimetre, and the temperature to a tenth of a degreeCentigrade. Next, introduce into the laboratoryvessel a few drops of a cold saturated solution of

caustic potash, and drive the gas back again intothe laboratory vessel to absorb C0 2 . In aboutthree or five minutes this will be effected, and thegas is returned to the measuring tube and againmeasured as before, except that it is not necessaryto observe the barometer again.

Now add to the contents of the laboratory vessela few drops of a cold saturated solution of pyrogallicacid, and return the gas upon it. When the quan-tity of gas is very small, and the vacuum thereforebut slightly impaired during the combustion, thereis occasionally given off a trace of oxygen, whichwill be indicated by a dark stain on the sides of thelaboratory vessel when the gas encounters the alka-line pyrogallate. If oxygen be present, allow thegas to remain for five minutes to absorb it. Thepresence of oxygen involves, of course, the loss ofany nitric oxide which has also escaped the reducingaction of the copper; but this is a very smallproportion of an already very small quantity, andtherefore is of no moment. If oxygen be notpresent, add a few bubbles of pure oxygen from apipette, a very convenient form of which is shownin Fig. 8, where a and b are two small glass bulbs

Fig- 8.

blown on a piece of glass tube, similar to that usedfor the Sprengel pump, about 50 m.m. in diameter.The connecting tube is constricted at c, so thatmercury can only pass very slowly through it; thedelivery tube, d, is drawn out to a capillary bore.Pure oxygen is obtained by heating dry powderedchlorate of potash in a bulb retort, made by blowinga bulb on a piece of combustion tube, drawing outclose to the bulb a very long narrow neck, whichafter the introduction of the chlorate is bent twicein opposite directions, so as to form a convenientdelivery tube, the sharply bent up end of which justdips below the surface of mercury in a mercurytrough. The first 60 or 80 c.c. of oxygen beingrejected, a sufficient quantity is collected in testtubes, each of which is removed from the troughin a very small beaker, containing enough mercuryto effectually confine the oxygen. The first tubeshould he tested by transferring about 10 c.c. of thegas to another tube, and heating it with alkalinepyrogallate. If only a very small bubble is leftunabsorbed, the gas in that and the succeedingtubes may be considered pure, otherwise the secondtube is similarly tested, and so on. The oxygenpipette is charged by filling the bulb, J, and thetubes, d and c, with mercury, introducing d into atest tube of oxygen, and drawing air from e; thegas will then enter b. It is necessary to see that dis filled with mercury before removing the pipette