218 DKSCRIPTTOW OF \ Poh/rffifff'rif'a. 



naturalist in this department has far sui'passed those of any european 

 observer, if we except Ehrenberg, (See section Navkulaceci). 



The composition of the lorica or shell in this family is various, 

 and may be separated into tAvo kinds ; the fii-st containing silica, 

 either pure or in combination with the oxide of iron ; forming a 

 silicate of iron ; the second, those in which silica is entirely absent, 

 when the lorica has a membranous or parchment-like (structure) 

 texture. It is remarkable that in no case has lime been foimd to 

 enter into their composition. In some genera the lorica is surrounded 

 by a soft gelatinous variously-fonned envelope or induvium. The 

 shape of the lorica is various, but such as entu'ely to inclose the in- 

 ternal organic portion, except the parts where it is united with 

 others ; hence it is termed urceolate. The lorica is composed of two 

 or more shells, or pieces, termed valves, which are usually dish or 

 cup-shaped, and often fluted (striated) or grooved. Those in which 

 silica enters iuto the composition have usually a round or a prismatic 

 four- sided figure, while, in the non- siliceous, they are generally flat, 

 with three to five sides. 



Of the internal organization of these creatures little is known, owing 

 to their opacity, and the structiu'e of the enveloping lorica. In many, 

 however, lai'ge ti'ansparent variable vesicles are seen among the mass 

 of coloured granules which occupy the greater part of the lorica. 

 These vesicles are considered by Ehrenberg to be digestive cells, and 

 the coloiu'ed mass (the ehlorophyl of botanists,) ova. In some 

 species, as soon as the coloured ova are protruded, the parent dies ; 

 in others, the ova form a Monad-like mass, fi'om which when ma- 

 tiu'ed the parent separates ; hence, says Ehrenberg, has arisen the 

 opinion of the transition of animals into plants. In Micrasterias, 

 Anthrodesmus, and one or two other genera, says Ehrenberg, male 

 reproductive structiu-es are visible, but no trace of a sensitive system 

 has been discovered. 



From the clustering nature of this sluggish family, and the rigidity 

 of thcu' coveiings, they resemble the confervoid Alga?, and other 

 minute vegetable fonns, and are hence confounded with them ; but 

 in their mode of propagation a distinction may be recognized. In 

 the Bacillaria, the self- division is always longitudinal, so that the 



