TRANSACTIONS OP SECTION D. 699 



Yellow and rod frequently follow the same rules of distribution as white. They 

 are well known to be fat-pigments. 



The author carefully guards hhnself against the extension of his theory to all 

 cases where white occurs in vertebrates. It is obvious that not ali animals are 

 subject to this atrophy, and there must be other causes for absence of pigment. It 

 seems highly probable from what the author has written that the known uneven- 

 ness of animal coloration is hut the external indication of uneven nutrition in 

 different regions of tlie body. 



2. A New Form of Osmometer for Direct Determinations of Osmotic Pres- 

 sure of Colloids. % Professor Benjamin Moore, if. ^., D. 5c. 



_ This form of osmometer has been specially designed to avoid leakage and pro- 

 vide a large surface for diffusion compared to the volume of solution employed, so 

 that the iniiuence of crystalloids is made slight and transient. To effect tlie iirst 

 purpose the instrument is made in metal in two halites which can be tightly 

 screwed together by a collar, and the two halves are further made up of'' two 

 shallow cells, to hold the solution and solute, so that the influence of admixed 

 crystalloids is rapidly eliminated by diffusion. 



The instrument essentially consists of two flattened circular cells of platinum • 

 5 cm. ni diameter and 1 cm. in depth, with flat flanges at their rims which tit 

 into corresponding thick flanged cases of silver-plated brass. The brass cases can 

 be screwed tightly together by a brass collar with a female screw at one edge, which 

 engages with a male screw on one of the brass cases, and a flange on the other ed^e 

 which catches on the flange of the other brass case. ° 



A diaphragm of platinum bored closely with holes about 3 mm. in diameter is 

 placed between the two cells, and serves to support the membrane of parchment 

 paper which separates them. 



The instrument is made pressure-tight bv two rings of thin indiarubbcr sheetintr 

 placed on either side of the diaphragm, and when the collar "is securely screwed 

 home has in all cases been found free from any trace of leakao-e. 



A platinum tube of about 4 mm. diameter leads, at right angles, from the 

 centre of the back of each cell, and is used to connect up for filling and reo-isterino- 

 the pressure. o o o 



The large surface provided for diffusion compared with the volume of solution 

 assures an early equilibrium of crystalloids. When a 1 per cent, solution of 

 sodiuni chloride is placed on one side and distilled water on the other, there is no 

 appreciable movement in the attached mercurial monometer, and within twenty- 

 four hours the amount of sodium chloride on the two sides is equal within the 

 limits of experimental error. 



The connecting tube on the side of the colloidal solution is joined up by rubber 

 tubing to a T-piece, which is in communication by one arm with a glass funnel and 

 by the other with a mercurial monometer. On the side of the solute (water) the 

 cell IS simply joined by rubber tubing to a funnel. 



In using the apparatus the connections are first filled with the respective 

 fluids and temporarily clipped off, then the cells are filled in turn and joined up, 

 all air being expelled by pressure upon the rubber connections. 



The two funnels upon either side are fixed at an equal height, of about 20 

 centimetres in each case, above the instrument, so as to ensure initial equality of 

 pressure on the two sides of the membrane. 



The osmometer is then suspended in a large vessel of water for the purpose of 

 maintaining a constant temperature, which can be arranged at different levels bv a 

 thermo-regulator. 



When the desired temperature has been attained, the osmometer is clipped off 

 from the funnel connection on the solution side and left in communication with 



' In earlier experiments the platinum lining was dispensed with and the silver- 

 plated brass case only used. This makes the instrument much less expensive, and is 

 effectual unless when the action of reagents, such as alkaUes and acids, upon the 

 colloid is to be tested. 



