516 Transactions of the Society. 



the slide; consequently the sections are not all perfect, for a 

 medium-size valve often contains more than sixty rows of dots. Be 

 this as it may, several sections are excellent. From this I infer 

 that the "lyra" figure is produced by thickened and chamberless 

 portions of the cell-wall. In like manner, the central nodule is a 

 large flat thickening of the wall. Fig. 9 probably shows the sec- 

 tion exactly through the middle of the valve (No. 13 of the series) ; 

 fig. lU section not far from the middle (No. 17 of the series) ; 

 fig. 1 1 section not far from the end (No. 1 of the series). In the 

 two last we see the thickenings, which I have designated as " lyra 

 plates," clearly project inwards. The sculpture of the dotted portion 

 of the valve may really be regarded as similar to Fleurosigma ; 

 but here are clear rows of isolated chambers closed all round. 

 Towards the edge the valves become thinner and the chambers 

 smaller. The midrib with the chambers adjacent to it can only 

 be seen faintly on most sections, especially the projection inwards 

 is seldom distinct, and in the first section, fig. 11, is not seen. 

 Longitudinal sections were not made ; they probably would illus- 

 trate the details more beautifully. Combining with these results 

 the surface view of a valve, fig. 12, we arrive at the conclusion 

 that the doubts I formerly entertained (6, pp. 482-4) with regard 

 to the existence of closed chambers, and of double membranes 

 connected by column-like supports, are unfounded. The clearly 

 separated spherules of the surface image also show separated 

 chambers, and this is equally true of the entire series of rows. 

 The space between such rows of dots, which is not rarely twice the 

 breadth of the chamber diameter, represents without doubt the 

 solid wall, which has not suffered a visible separation ; hence there 

 is no communication between the separate rows. If we have thus 

 before us the connecting link to Fleurosigma, it only requires 

 another step to arrive at Finnularia : if the chambers forming one 

 row are brought together a little closer and coalesce with each 

 other, then we get the extended cylindrical chamber of the former. 

 It is true that the large opening is unconnected. 



3. Fleurosigma. 



§ 1. Addenda to my former researches. — I add a few recently 

 obtained results, chiefly due to the method subsequently learnt of 

 placing the frustules in position and making serial sections (vide 

 supra, " Method of Investigation," (1). 



(1) I was formerly obliged to neglect the transverse section of 

 the central nodule (6, p. 478), because I could not find it ; but in 

 serial preparations of P. halticum as well as P. angulatum, cut 

 exactly transversely through a bundle, there is no difficulty in 



