51 



Baden, French Lick, Mt. Aris, Indian and Trinity Springs. The results 

 were constant and accurate. It also holds in waters containing much mineral 

 matter and organic matter to 350 parts in 100,000. 



Second. "Will there, without filtering, be an appreciable error? We pro- 

 duce the following data : 



WITH DISTILLED WATER. 



We differed from the authors in this — that the solution was gently agitated 

 until the color no longer disappeared. With such a standard the error may 

 reach 1.77 pts. in 100,000 pts. as shown by the eighth titration. The observa- 

 tions justify the following inferences : 



First. There is an appreciable reaction between the silver chloride and 

 the ammonium sulphocyanide. 



Second. The error varies directly with the quantity of chlorine present, 

 and the time employed. 



Third. When the reaction of chlorine upon silver nitrate is effected in 

 the presence of ammonium sulphocyanide, the results are inconstant. 



Fourth. That it is necessary to filter off the silver chloride, before add- 

 ing the ferric salt. 



Fifth. That by filtering the results are eery accurate. 



Sixth. That if the solution, unfiltered, be allowed to stand ten minutes, 

 the reactions which take place, will produce very appreciable errors. 



Some suggestions to teachers of science or mathematics ix high schools. 

 By Thos. C. Van Nuys. 



It is the purpose of the writer to endeavor to indicate, as briefly as prac- 

 ticable, the spirit which should influence teachers of science or mathematics 

 in high schools. » 



It is needless to state in this connection that the spirit, in which a teacher 



