1901.] NATURAL SCIENCES OF PHILADELPHIA. 321 



holding two or three litres is employed and into this a considerable 

 quantity of Cabomba Carolinlana or Myrio2:>hyllam sjncatuvi is 

 introduced (fig. 2). The water is thoroughly charged Avith carbon 

 dioxide and the plants are then exposed to the sunlight. Little 

 streams of gas are seen to pass upward from various points, and 

 when sufficient gas has collected at the top of the flask, the latter 

 is immersed in a tank of water in a horizontal position in such a 

 manner that the gas is directly under the opening (fig. 3). On 

 turning the stop-cock and applying a splinter of wood with a 

 spark on the end of it the gas will be found to be oxygen. 



When the supply of carbon dioxide in the water has been 

 exhausted the plant will no longer give off bubbles of oxygen. 

 The process may be again initiated by passing carbon dioxide into 

 the receiver. Before testing it is best to allow the carbon dioxide 

 to become exhausted, since in recharging the water it is impossible 

 to avoid collecting some of this gas over the liquid and adulter- 

 ating the oxygen. On standing it is gradually absorljed by the 

 water and consumed by the plant. In any case the gas collected is 

 not pure oxygen, but it is sufficiently rich in this substance to 

 make an eflectiv'e demonstration. 



The deaths of D. Shepherd Holman, a member. May 13, and 

 of Thomas C. Porter, a correspondent, April 27, were announced. 



May 21. 



Mr. Charles INIorris in the Chair. 



Seventeen persons present. 



Papers under the following titles were presented for publication: 

 " Fishes from the Caroline Islands," by Henry W. Fowler. 

 " Types of Fishes," by Henry W. Fowler. 



Strudure of Diatoms. — Mr. Frank J. Keeley remarked that 

 in studying ihe structure of diatom valves some years ago the 

 method employed : mounting broken valves at right angles to the 

 cover glass, proved efficient for most of the coarsely marked forms, 

 but failed with certain species of Aulacodlscus. 



Such forms as ^-1. SoUittianns, A. uiargarataceous, etc., yielded 

 satisfactory sectional views and proved not to differ materially in 

 structure from Coscmodiscus ; but another group, including A. 

 Oreganus, A. Rogersii, A. Jaaischii, etc., proved too opaque for 

 the elucidation of their structure by this method. Further exam- 



21 



