Mr. J. Ralfs on the Diatomacere. 347 



In the first section, the species belonging to which are gene- 

 rally marine, the ends of the frustules are convex, and as there 

 are no distinct junction-surfaces, a nioniliform appearance is pro- 

 duced. The central line is more strongly marked in this than in 

 the other section, and seems to divide the frustule into two equal 

 portions. It becomes broader and then double, and at length an 

 intermediate growth separates the two halves of the frustule, which 

 dm-ing this process do not increase in size ; but w hen the interme- 

 diate space is equal to the diameter of the original frustule, two 

 new frustules are formed by the addition of two hemispheres on 

 the inner sides of the separated portions. The outer siliceous 

 covering still remaining, the frustules are connected in pairs, and 

 appear like two globules within a joint, as they are characterized 

 by Harvey in M. nummuloides, and by Carmichael in M. glohifera. 

 The above description belongs more particvdarly to M. nummu- 

 loides, but the process in the other species in the first section is the 

 same. Mr. Dalrymple first called my attention to this singular 

 mode of growth. In a letter he remarks on M. nummuloides, that 

 " the joints of the Meloseira are not only united end to end, but 

 sometimes two corpuscles are included within the same boundary- 

 line, while others are disjointed as it were and merely united end 

 to end. Tliis I believe arises from the mode of reproduction, and 

 that each corpuscle is capable of dividing into two bodies within 

 its original siliceous shell." A series of changes nearly similar 

 in Isthmia has been abeady described*. 



All the species in the other section are found in fresh water. 

 The frustules are not united in pairs, their junction-surfaces arc 

 distinct and nearly flat, and their central lines are probably fur- 

 rows ; hence whatever parts of these lines occupy the margin of 

 the field of \dew, as the filaments are tm-ned round, they all ap- 

 pear like puncta. 



M. lentigera, Harv., evidently belongs to the first section, but 

 I have seen no specimen, and cannot refer it to any species with 

 which I am acquainted. 



Through the kindness of Mr. W. Thompson I have been en- 

 abled to examine a specimen of " Meloseira Thompsoni " marked 

 by Mr. Harvey himself. This beautiful and interesting plant ap- 

 pears to me to be a species of Sph(Efoplea ; at any rate it should 

 be removed from Meloseira, since the filaments are not siliceous. 



Mr. Hassall has favoured me with the followdng observations 

 on this plant : — " I can scarcely believe that Meloseira T/unnpsoni 

 is really a Meloseira, for it does not appear to exhibit even one 

 of the characters of that genus. The filaments in M. Thompsoni 

 are equal, while in the species with which I am acquainted of true 



* This similarity has also been observed by Mr. Berkeley. 



