DIATOM LOCOMOTION. 1 55 



flask at the place where the little opening in the paper had 

 been. This deposit was found to consist of quite pure A^. 

 Iridis. Under proper precautions the water of the flask was 

 examined and no diatoms were found freely floating. The 

 walls elsewhere were lightl}' strewn with frustules, driven by 

 light-hunger up the overhanging, conical walls of the flask to 

 the little ray of light in the darkness. 



If the diatoms are capable of so much, why should they 

 not succeed in moving perfectly well upon the under surface 

 of the cover glass ? It is only one step farther. To test the 

 matter for various species, I made use of a very shallow cell. 

 This cell is constructed of two parallel strips of thin glass, 

 cemented upon an ordinary glass slip. The cover glass rests 

 upon these two strips, and the cell is open on two sides and 

 so shallow that a 1/5 inch objective will readily focus down to 

 the bottom. A diagram of a section of the cell is given oppo- 

 site, A being the glass slip, B B the two thin strips (greatly 

 exaggerated in relation to thickness) and C the cover glass. 

 In order to use the cell for the present purpose, the diatoms 

 to be observed are dropped in with enough water to fill the 

 space, the cover glass is placed, and the whole arrangement 

 reversed so that the cover glass is underneath and the slip 

 above. As soon as the diatoms have settled, the cell is 

 brought carefully into normal position with C on top. Manj- 

 frustules will be found clinging to the under surface, and 

 soon they will begin to move. 



Navicida Iridis is represented in median cross-section at 2 

 in the diagram. This diatom, unlike many of its genus, has 

 a very thin coleoderm, so that the rather flat surface of the 

 valve near the middle of the length appears to be in close 

 contact with the glass. The raphe and the central nodule are 

 slightly iDelow the general beaded surface, but the former is 

 still almost in contact. Toward the pointed ends, the surface 

 curves slightly downward, like the keel of a flat-bottomed 

 boat at bow and stern. In addition to the position illus- 

 trated, where mere "adhesion" between the flat valve and 



