BROMINE IN SEA WATER. 89 



the contents of the graduated cup nearly as light as those 

 of the other cup, open the clamp, mix and wash the 

 cups as described above. As I know that I am very 

 near the finish, I raise only to the 1 c.c. mark. 



No. 2.— I add one drop: the 1 c.c. gets qnite light 

 compared with its partner in the other cup. I blow, 

 now, into my rubber tube, open the clamp while blow- 

 ing, in order to prevent the sample from sinking back 

 into the flask, and raise to the 5 c.c. mark. This 

 permits a better observation than adding the drop at once 

 to the 5 c.c. I compare the two colors at my leisure. 

 The result is doubtful. The color seems to have neither 

 increased nor decreased. The next drop shows a decided 

 decrease of color on comparing both cups. I cannot 

 go any further with fractitional titration, being, prob- 

 ably, within three drops of the end of the assay. I open 

 the clamp, remove the cork and funnels after washing 

 rod and cups with water, and boil the contents of the flask 

 until the color has become quite white and no more smell 

 of bromine can be noticed. I cool by water, add one c.c. 

 of chloroform, close the flask with my thumb, shake vio- 

 lently and observe whether the chloroform has become col- 

 ored. If not, as it will not be if all the bromine has been 

 evaporated, I add one drop from the burette and shake 

 as before. In the settled little drops of chloroform (the 

 advantage of a pointed egg-shape of the flask will show now) 

 I recognize distinctly a yellowish tint. The next drop leaves 

 me uncertain if there is any change in color. The third 

 drop shows without fail an almost complete decoloration. 

 The assay is finished. By deducting the last drop and one- 

 half of the former drop, I learn that 25 milligrm. bromine 

 are indicated by 3.925 c.c. of my standard solution. Each 

 drop of .05c.c. will, henceforth, disclose (^;^|^ being =78.5) 

 ^\% or about ^ of a milligram of bromine. The limits of 

 accuracy depend of: (1) the least possible number of drops 

 for testing by chloroform; (2) the least possible amount of 



