164 JOURNAL OF THE [December, 



in consequence of having exhausted the silica in the water. So 

 I submitted four one-pint bottles of the water to the skillful an- 

 alyst Prof. Peter T. Austen, of Rutgers College, New Brunswick, 

 N. J. These bottles were labelled A, B, C, and Z>, respectively. 

 The first two contained water that had not been used in my ex- 

 periments, A being from the large demijohn, B from the smaller 

 one, the water in both having come from the same place in 

 Raritan Bay ; but while the water in the smaller demijohn had 

 been in my cellar not quite two years, that in the larger one 

 had been there nearly seventeen years. Samples C and Z> were 

 from the jars of the first series of experiments, which had been 

 so fruitful of diatoms. To my surprise, the following was the 

 report of the analyst : — 



Amount of Silica in grains per gallon. 

 Sample A, 1.7 1 

 Sample B, 1.24 

 Sample C, 3.78 

 Sample D, 4.44 



The Professor wrote me : "You will notice that the samples 

 in which the samples have been grown are much richer in silica 

 than the original water. This is probably because the water has 

 stood for a long time in glass vessels and has hence dissolved 

 silicates from the glass. It is probable that this silica is com- 

 bined as silicate of soda or potash. Water has a very appreci- 

 able solvent action on glass, indeed, so much that in chemical 

 analysis we have to correct the error introduced by the use of 

 glass vessels. Without large samples and exhaustive analyses 

 it is difficult to interpret these results. A possible hypothesis 

 for the refusal of the diatoms to grow any longer may be that 

 they only take up free (that is) uncombined silica, or, if they 

 take the silica away from the base, the setting free of the base 

 may leave the water more or less alkaline, which may possibly 

 be detrimental to their development. It is very difficult in a 

 case like this to get all the conditions." 



It should be noted that the unused water in samples A and B 

 had, so far as being in glass vessels is concerned, been all the 

 time in darkness ; also, that A, from the larger demijohn, is richer 

 in silica than is B. But C and B> had been exposed nearly 

 five years to daylight; hence, in my opinion, the solvent action of 



