142 Mr. Waldie's Investigations connected [No. 2, 



General's Tank water, 0385 



Dalhousie Sq. Tank water, 0265 



Racny Season. 



River water of lOtli July, 0240 



„ „ „ 13th July, 0810 



„ „ „ 16th August, 0380 



„ „ „ 24th August, .0140 



Average, 0392 



These results exhibit great variations in the proportion of ammonia, 

 and recent investigations throw some doubt on the trustworthiness of 

 the method employed ; besides, it is confessedly an imperfect indication 

 of the amount of nitrogenous matter : nevertheless, so far as they go, 

 they are confirmatory of those already given in table IV. of Part I. 

 The waters of the rainy season yield more ammonia than do those 

 of the hot season, and the tank waters, even the best of them, 

 yield more than the river water. And the previous table also shews 

 that General's Tank, — probably the best of the tanks, or at least 

 one of the best, contains more organic matter than the river water ; 

 and the results given connected with Dalhousie Square Tank confirm 

 the opinion expressed before, that the river water rather deteriorates 

 by storage in tanks. 



I do not mean to say that these results, as to the quantity of 

 organic matter and the proportion of nitrogenous matter, prove with 

 certainty that the river water is at least equal, if not superior, to 

 even the best of the tank waters ; but they are the best means of 

 judging which chemical analysis affords of the quality of water for 

 drinking purposes, and, in the absence of equally good evidence to the 

 contrary, the results lead to such conclusions. 



Only in one point have I to say something in modification of the 

 results of previous examination , and that is connected with the water 

 of the rainy season. In the previous pages I have spoken strongly 

 of the putrid flavour of the river water of the rains, particularly of 

 the early part of the season. This respected the year 1866. In the 

 present year, however, I have not found this putrid flavour, or at least 



