NATURAL HISTORY OF THE DIATOMACE^E. 443 



by the frustules sliding back again, as it were, over each other ; and this changed 

 direction of movement proceeding, the filament is again drawn out until the frustules 

 are again only slightly in contact. The direction of the movement is again reversed, 

 and continues to operate in opposite directions, the time occupied in passing from the 

 elongation in one direction to the opposite being generally about forty-five seconds. 

 If a filament while in motion be forcibly divided, the uninjured frustules of each por- 

 tion continue to move as before, proving that the filament is a compound structure, 

 notwithstanding that its frustules move in unison. When the filament is elongated to 

 its utmost extent, it is extremely rigid, and requires some comparatively considerable 

 force to bend it, the whole filament moving out of the way of any obstacle rather than 

 bending or separating at the joints. 



This' is not always the case, as I have myself frequently observed, as 

 the filament often becomes bent by the force of its own motion. And 

 there is one other fact which seems to have escaped the notice of Mr. 

 Thwaites, and which adds considerably to the interest of an exhibition 

 of this plant while in motion. After the extended frustules have returned 

 to their normal position in the filament, so as to form a ribbon-like com- 

 bination again, and just at the time when they are about to start on their 

 way in the opposite direction, there seems as if a considerable amount of 

 force were necessary to get them started past this point, for this force is 

 very apt to dislodge the whole filament and swing it entirely round, so 

 that that end which was, we may say, pointing towards the right hand, 

 now points to the left. The consequence is, that now, when the frustules 

 proceed on their path towards the opposite side of the filament from 

 which they projected previously, they, in reality, extend towards the same 

 side of the microscope as they returned from. This motion often contin- 

 ues to be exerted, so that the whole filament is periodically swung around 

 on its centre as the frustules return to their places, and become again 

 parallel one to the other. Often and often have I spent hours looking at 

 this marvel of nature; the motion without apparent cause or mode, an 

 invisible joint, which, as a friend of mine (an engineer) once remarked, 

 would be a fortune to any one who would discover it, for here we have 

 several sticks forming the bundle, moving over each other without sep- 

 arating, and yet the use of the highest powers of the microscope has 

 failed to detect the means of their union into one mass, or composite 

 group, of individuals. The more often I watch Bacilleria paradoxa, the 

 more it puzzles me. Not long since I saw one specimen (of course, I 



