December 16, 1921] 



SCIENCE 



609 



uation is completed througli the fine stage, 

 B. In this unit a portion of the high-pres- 

 sure vapor from the central tube is allowed 

 to expand to a low pressure through one or 

 two small openings into the inverted cup, C. 

 This vapor then escapes freely into the large 

 water jacketed tube and gives the conditions 

 essential for high-speed exhaustion. 



Tt has found that the high-pressure stage 

 operating alone, without assistance from the 

 low-pressure unit, will produce a high vacuum. 

 The speed of the high-pressure unit by itself, 

 however, is very much less than that of the 

 combination, which possesses a speed compar- 

 able with that of a single stage pump of 

 equivalent proportions. 



The advantage of the combined units, of 

 course, lies in the fact that such a pump 

 will function in a perfectly satisfactory fash- 

 ion with a very ordinary fore-vacuum. A 

 mechanical pump capable of reducing the 

 pressure to 2 or 3 millimeters is satisfactory, 

 or even a water aspirator which will give a 

 vacuum of 20 millimeters can be used if 

 nothing better is available. 



"With regard to the construction of the pump 

 perhaps a little may be said. Glass posses- 

 sing a low coeificient of expansion such as 

 Pyrex or Corning G702P glass must be used 

 in making it, as otherwise one will almost 

 certainly experience the rather annoying in- 

 convenience of having the boiler crack upon 

 application of the heat. The size of the pump 

 can, of course, be varied considerably, but 

 the general proportions of the parts given in 

 the drawing are found to be very satisfactory. 

 In the pump from which the drawing was 

 made the mecury boiler has a diameter of 90 

 millimeters and the other dimensions were 

 rediiced proportionately. The dimensions of 

 the jet and throat which have been found to 

 work well are indicated in the enlarged sketch 

 of this part. The diameters given apply to the 

 tube openings. The thickness of the nozzle 

 wall should be as thin as is consistent with 

 reasonable strength. The two small open- 

 ings which serve to furnish a supply of vapor 

 to the upper unit are about the size of ordi- 



nary pin holes and are located on opposite 

 sides of a small enlargement in the central 

 tube. The joint between the lower end of the 

 water jacket and the body of the pump is 

 made water tight by binding it tightly with 

 strips of thin rubber. There is some advan- 

 tage in having a slight constriction where 

 the mercury return tube is sealed to the boiler 

 as the presence of a constriction here tends 

 to preserve the equilibrium of the mercury 

 in the return tube. 



The mercury in the boiler should be about 

 2 centimeters in depth at the center and ord- 

 inarily, with a properly adjusted flame, it 

 will evaporate without serious bumping even 

 at the higher pressures. The height of the 

 mercury column in the return tube indicates 

 the vapor pressure in the boiler and the pres- 

 sure required for satisfactory pumping de- 

 pends entirely upon the fore-vacuum. There 

 is no harm, however, in running the vapor 

 pressure up as high as the length of the re- 

 turn tube will permit if this be necessary to 

 enable the pump to function. 



E. H. KURTH 



Palmer Physical Laboratory, 

 Princeton, N. J. 



THE AMERICAN CHEMICAL SOCIETY 



(Continued) 



DIVISION OF AGRICULTURAL AND FOOD CHEMISTRY 



C. E. Coates, Chairman 

 T. J. Bryan, Secretary 



The testing mid grading of food gelatins: 

 Clarke E. Davis and Earl T. Oakes. Loeb 's re- 

 cent work on gelatin is briefly discussed and Ban- 

 croft's objections to Loeb's conclusions on the 

 basis of the insolubility of gelatin as based on sur- 

 face tension measurements by Slobeki are shown to 

 be in error. Methods for determining gel strength 

 and viscosity are given and the effects of various 

 factors affecting these properties are discussed with 

 data. Data on the causes for discrepancies between 

 grading gelatins by gel strength tests and by vis- 

 cosity measurements are given. Gelatins submitted 

 by the manufacturers as examples in which gel 

 strength does not parallel viscosity are shown to be 

 classified alike by gel strength and viscosity meas- 

 urements under the methods described. 



Active chlorine as a germicide for milk and 



