342 THE AMERICAN MONTHLY [Nov 



biology is developing nambers of well-trained observers 

 wide awake to the value of these problems and well able 

 to undertake tlie work. What is needed is cooperation. 



Various methods have been employed from time to time 

 for determining the character and amount of microscopic 

 life in water. Those interested in the subject from the 

 piscatorial standpoint have usually employed some sort 

 of net for straining the organisms from the water and 

 concentrating them for the microscope. One of the best 

 devices of this kind is that devised by Professor Reighard 

 and used with good results for studying the plankton in 

 Lake Michigan. It consists of a conical net of fine bolt- 

 ing cloth, at the small end of which there is a 'bucket/ 

 made by covering a metal framework with some of the 

 same bolting cloth. The apparatus is hauled through 

 the water, filtering a column of water whose cross section 

 is the same as the circular mouth of the net and whose 

 length is equal to the distance through which the net is 

 hauled. The organisms are caught by the fine bolting 

 cloth and are ultimately washed into the bucket. The 

 collected material is then removed by an ingenious 

 arrangement, measured and sent to the laboratory for 

 microscopical examination. By this method one is ena- 

 bled to get a good idea of the total amount of suspended 

 matter in the water, but it can hardly be called an accu- 

 rate method of obtaining the number of living organisms 

 present, as the net sweeps in amorphous matter as well 

 as organisms and some of the smaller forms undoubtedly 

 escape through the bolting cloth. Moreover, the amount 

 of water actually filtered cannot be told with a great 

 degree of accuracy. Nevertheless, the method is one of 

 value, particularly for securing the larger and rarer 

 forms of rotifers, Crustacea, etc. 



Sanitarians who have studied the microscopical organ- 

 isms in water supplies have usually employed very dif- 

 ferent methods from the above, partly because they have 



