On estimating Ammonia in Potable Waters. 257 



these three is provided with a glass tap for filling and emptying 

 the apparatus.* 



The water to be examined is poured into the retort, and one- 

 third distilled over gently, but at a fairly good rate. If operating 

 upon half-a-pint, the operation should take about one hour. The 

 whole apparatus is previously connected together, and the two 

 bulbs are filled with water free from ammonia (i.e., water which 

 has been boiled rapidly in a flask, until it gives no reaction with 

 the Nessler reagent). The bulbs are easily filled by opening 

 the tap, inserting the point of the tap tube into a beaker of the 

 water, and sucking with the breath gently — at the same time, 

 stopping the other end of the bulb tube by the finger. 



After the operation is completed, you have the most of the 

 ammonia, if not all, in the receiver, whilst any vapour which 

 might have escaped therefrom is caught in the tube " A ; " at the 

 same time, any atmospheric ammonia which enters by regurgita- 

 tion is caught in the bulb tube '•' Z." " A " is run into the dis- 

 tillate in the receiver, whilst "Z" is rejected. In practice, it is 

 not necessary that the water in "Z" should be changed for twenty 

 or thirty operations, providing that an increased amount of am- 

 monia is not about in the laboratory atmosphere ; but if so, it is 

 desirable to change with each experiment. 



If we consider the points that I have dwelt upon as being 

 sources of error, we shall see that this simple apparatus fulfils 

 absolutely the necessary conditions, and that the tube connecting 

 the bulbs "A" and " Z," contains an atmosphere perfectly neutral 

 as regards ammonia, because it is the space between two liquids 

 which do not give off" ammoniacal gas, and which give off vapour 

 of an equal tension. 



I may mention that, on submitting it to experimental proof, 

 the use of an acid in either of the bulb tubes proved a mistake. 



It is not necessary to give further details, but I may as well 

 slate that in many crucial experiments, I have been in the habit 

 of estimating the amount of ammonia in the receiver, and both 

 bulbs "A" and "Z." The following, as regards Vartry water, will 

 serve as an illustration : — 



Distillate in receiver, .... 0-003 per 100,000. 

 Bulb "A" (nearly), . . . . O'OOl „ 



Bulb " Z," gave an indication which affected the Nessler, but 

 which was not sufficient to estimate. 



* Bulbs of ray pattern are made by Messrs. Cette & Co., of London. 



