EDWIN JOSEPH COHN 711 



the neck of the wide mouthed bottle that is used is always in the same position, 

 and a pipette may be inserted and a sample withdrawn even when the bottle is 

 performing more than a revolution a second. The base of the carriage is attached 

 through a sliding valve (£) and a universal joint (F) to an eccentric (G) on a driving 

 wheel (H). Instead of being fixed at a definite distance from the centre of the 

 wheel, (in which case the motion of the bottle would be circular) the eccentric 

 moves freely on ball-bearings in a slot (7) but is forced by a spring (J) to impinge 

 on a track (K) with which it makes contact through a roller-bearing (L). The 

 shape of the track alone determines the motion of the carriage. An ovoidal 

 shape has been adopted in order at once to prevent foaming and centrifugal action. 

 With the aid of this apparatus we have been able to sample protein suspensions 

 with an accuracy of 0.3 per cent. 



After the protein had been delivered into the volumetric flasks the 

 solvent was slowly added, drop by drop, from a burette until it rose 

 to the graduated mark on the neck of the bottle. In certain experi- 

 ments, the space above was further reduced by the addition of glass 

 beads of a grade that gave off no measurable alkali, and by a drop of 

 toluene added to prevent bacterial action. The flasks were finally 

 stoppered, fixed in place in the carriage of a specially designed shaking 

 machine (Fig. 2), and shaken in a water thermostat at 25.0° ± 0.1°C. 

 WTiile the shaking machine was in motion, the necks of the flasks were 

 covered with inverted test-tubes to prevent contamination from the 

 spray of the bath. 



As the carriage went back and forth on tracks in the water bath, 

 the glass beads in the small volumetric flasks functioned as a very 

 efficient ball mill and kept the protein precipitate mechanically sus- 

 pended. In our experience the heavy flocculent isoelectric protein 

 has usually been brought into equilibrium with its solvent in from 

 24 to 72 hours. The shaking machine was then stopped, the contents 

 of the flasks allowed to settle slightly, and the saturated solutions 

 filtered on No. 42 WTiatman filter papers. By means of a water jacket 

 through which the water of the thermostat was kept in constant 

 circulation by a rotary pump, the filtration was carried on at nearly 

 the same temperature as the bath. At least 8 cc. of the protein 

 solution was always used to wash the filter paper, and was accordingly 

 discarded. Aliquot parts of the remainder of the filtrate were meas- 

 ured out with carefully calibrated pipettes into Kjehldahl flasks. 

 Analyses were usually made in triplicate. 20 cc. of sulfuric acid 



