90 THE AMERICAN MONTHLY [Mar 



shine soon turned pink and afterwards purple. I had 

 seen this purple deposit covering considerable spaces in 

 the salt marshes and had tested it enough to find that it 

 contained a considerable amount of free sulphur, but 

 what gave it its photographic properties I did not make 

 out. This substance under the microscope appeared com- 

 posed of very minute particles, the larger of which were 

 of flattened ovoid form, and the smaller of irregular 

 shapes, and all in rapid motion. 



After several days exposure, the sides of this bowl be- 

 came covered simultaneously and uniformly with a light 

 brown film which the microscope showed to be com- 

 posed of countless myriads of a small organism resem- 

 bling miniature Nitzschia closterium but showing little 

 trace of its characteristic markings and being apparently 

 destitute of silex as they completely dissolved in boiling 

 water. Their color and their motions were those of dia- 

 toms, but otherwise they resembled desmids. 



I sent small vials filled with these organisms to many 

 experts, explaining the circumstances and asking their 

 opinions. I received various replies. Most of them 

 were in doubt, but Prof. H. L. Smith wrote that they 

 were certainly Nitzschia closterium, that he was very 

 glad I had them in such quantity and such purity, and 

 that he hoped I might succeed in keeping them long 

 enough to observe their method of conjugation which he 

 believed had never been seen in this species. 



Now, note here, that I had made numerous examina- 

 tions of this material and thought I had observed every 

 species it contained, but had not seen a single individual 

 of this species until the sudden appearance of this film. 

 I had little hope of being able to protect these diatoms 

 against their enemies for any great length of time, but to 

 take all the chances possible, I isolated portions of the 

 film in various small vials of different sizes, and made a 

 large number of life slides. The life slides proved a 



