OF ARTS AND SCIENCES. 87 



bar in Cocbituate water was brigbt for some montbs, but bas at length 

 become slightly dimmed in small patches, which may be attributed to 

 the less complete expulsion of the air by boiling, or the less accurate 

 stopping of the flask, though at the time the experiment was made both 

 were regarded as unobjectionable. 



" The following experiment shows how much is due to a change of 

 water. The bars in the Cochituate remained quite bright, and those in 

 the other waters were but slightly coated. Two bars in 15cc. for thir- 

 teen consecutive days, without changing the water, gave, in Cochituate, 

 0.500gr. ; Croton, O.SOOgr. ; Fairmount, 0.500gr. ; Jamaica, l.OOOgr. 



" These experiments seemed to show that, without a renewal of the 

 air, the action nearly or quite ceases after a short time. Professor Sil- 

 liman, Jr., made a similar observation in his experiments with the vari- 

 ous waters submitted to him for analysis by the Water-Commissioners 

 in 1845. He used a large volume of water, and yet the bar remained 

 quite bright. There was no alternate exposure to water and air. Chris- 

 tison remarks, that, while certain waters might doubtless be kept with 

 safety in leaden cisterns, the covers of the cisterns should not be of 

 lead, but of wood, since the moisture condensing on them, furnishing, 

 as he observes, pure water, would act on the lead, and the product fall- 

 ing would poison the water. The joint action of air and water is here 

 presented under exceedingly favorable circumstances. The corrosion 

 of cisterns along the line where air and water meet might be expected. 



" It will be readily seen, from considering the important part air 

 plays, how rain-water must act with great vigor upon lead. It contains 

 air, and is surrounded by air, and, aside from temperature, could not 

 be more favorably constituted for acting upon lead. The well-known 

 prevalence of lead maladies in Amsterdam, while leaden roofs were in 

 use, and the restoration of health on their replacement with tile, find 

 here a ready explanation. Dr. Dana has recorded an experiment with 

 rain-water, which furnishes a valuable confirmation of what is stated 

 above.* In a series of experiments with lead pipe of considerable 

 length, if an interval of half a minute, or even less, occurred between 

 the emptying of the pipe and refilling, there was invariably found lead 

 in the water. This has been observed on a large scale in the practical 

 service of lead pipe Where from any cause the pipes have been 

 empty for a length of time and then filled, the first water drawn con- 



* Appendix to Tanquerel. 



