2 Dr. J. D. Macdonald on the Structure of the 



He has also shown how the capacity of the frustule may be 

 augmented, at least in one direction, by the sliding out of those 

 plates or "rings" telescope-fashion to accommodate themselves 

 to the increase of the contents during division. This great fact 

 is barely shadowed forth in Griffith and Henfrey's ' Microgra- 

 phic Dictionary' (ed. 1856), p. 201, where it is said that u 'mBid- 

 dulpMa and Isthmia, and similar forms, the new half-frustules 

 formed inside the l hoop ' slip out from it like the inner tubes 

 from the outer case of a telescope." The " inner tubes " in 

 this case would of course be the hoops of the new valves, 

 which in their turn assist in forming a so-called " intermediate 

 piece " with their parent valves, though this is not specifically 

 stated, and, strangely enough, a different inference may be 

 drawn from the following quotation (op. cit. p. 199) : the frus- 

 tules of Diatomacese are described as consisting of " two usually 

 symmetrical portions or valves comparable to those of a bi- 

 valve shell, but are in contact at their margins with an inter- 

 mediate piece (the 'hoop'), variable in breadth, according to 

 age." This view of the structure of the frustule is substan- 

 tially the same as that given by Smith in his introduction to 

 the i Synopsis of the British Diatomacege,' only that he makes 

 the hypothetical intermediate piece or " hooj) " more insigni- 

 ficant by calling it a " connecting membrane," whereas it is 

 in reality double, as before stated, one portion being included 

 within the other, so as to admit of extension of the frustule in 

 the direction of the axis of growth. 



Each of these sliding segments, moreover, is not merely 

 connected but directly continuous with the body of its own 

 valve, that which is invaginated being always the younger, 

 having been produced within the other or the parent valve by 

 an endogenous process, combining fission with growth and 

 remodelling of the primordial utricle. Whatever be the con- 

 figuration of the true ends of the frustule, or, in other words, 

 the body of the valves, viz. circular, triangular, foursquare, 

 or navicular, the sides or "hoops" of the two forming the 

 so-called " intermediate piece" are, as it were, marginal exten- 

 sions of them, but perpendicular to their general plane. 

 Quoting, again, from the ' Micrographic Dictionary,' p. 200, 

 it is stated that " the ordinary mode of increase of the cells of 

 Diatomacese is, like that of other vegetable cells, a process of 

 division. ... It may be briefly described thus : — the pri- 

 mordial utricle, enclosing the contents, divided into two por- 

 tions, which separate from one another in a plane parallel 

 with the sides of the individual frustules ; the two valves of 

 the parent cell gradually separate from one another, remain- 

 ing connected by the simultaneous gradual widening of the 



