TERATOLOGY OF DIATOMS. 55 



the valves, and surrounded by a membrane which is primarily cellulose, 

 passes out into the cultivations, usually assuming a very irregular form, 

 but at the same time continuing to approach to that of normal mega- 

 frustules. Many of these bodies, however, only complete their ultimate 

 form when they begin to move. 



" 3rdly. The primordial mega-frustules of bizarre and unsymmetrical 

 shape acquire their elegant regularity by deduplication, a process to which 

 they immediately submit. 



" 4thly. Protoplasm which escapes from micro-frustules is not usually 

 rendered effective in the glutinous or gelatinous substance, as recorded 

 by various authors, and further, it is probable, as I have frequently observed, 

 that the double auxospores placed side by side are due to the simultaneous 

 germination of two micro- f rustules on the point of dividing, the internal 

 valves of which still adhere to another at certain points. 



"5thly. It is still uncertain whether micro-frustules before their germina- 

 tion are the subject of a special fecundation, the phenomena of 

 conjugation therefore must be for the present withdrawn as the correct 

 explanation of the cause of the re-establishment of form, in the case of 

 the five or six species on which I have been able to make observations." 



6. Teratology of Diatoms. 



Dr. Miquel in his articles on cultivations published in " Le Diatomiste" 

 expresses himself as follows : 



"In addition to normal cultivations others can be produced, in which 

 the predominence of certain physical and chemical elements can be in- 

 tentionally exaggerated; then when the diatoms are able to grow in these 

 media they acquire bizarre forms, which has induced me to call these 

 cultivations Teratological. I have been able to follow these strange 

 variations of form through three generations in certain Nitzschia and 

 Cydotella. Nothing is more curious than observing these original diatoms 

 getting embossed, asymmetrically attenuated and becoming absolutely un- 

 recognizable ; in the case of Cydotella, which are ordinarily in the shape 

 of a regular box, the valvular surfaces are seen to lose their circular form 

 become oval, triangular, quadrate, or assume the appearance of unangular 

 but very irregular closed curves ; at the same time the plane surfaces 

 of the discs are warped and become undulated, the edges of the upper 

 and lower surfaces of the cylinder begins, as it were, to form hills and 

 valleys, while the markings of the Cydotella are nevertheless persistent 

 and manifest themselves with these marvellous modifications superadded ; 

 on the girdle face the alterations of shape which I record are very 



