NOTICES OF BOOKS AND MEMOIRS. 219 



ill traiisverse section) four-sided and an outer three-sided cell. 

 The former of these cells is now divided by a pericliual wall into 

 an inner and an outer, and in this way a central square {Grimd- 

 quadrat) makes its appearance. This method of growth corresponds 

 closely with the process observed by the author himself in 

 JungenjiannicB, and by Kuhn in Andrecea. Each of the eight 

 peripheral cells is now segmented into an inner and an outer by a 

 periclinal, and each of the new outer ones by a radial wall. So 

 far the development of seta, apophysis, and capsule is uniform. 

 In the seta, the peripheral cells first divide sometimes by radial, 

 sometimes by periclinal, septa, and the same method of division is 

 sometimes relocated, so that some irregularity is seen in the dis- 

 position of these cells : meanwhile those of the central- square 

 become segmented in the same way as the primordial quadrants 

 were. In the apophysis the process in the central square is the 

 same, but in the peripheral cells, after the first radial septum, two 

 periclinal septa are laid down, after which the new peripheral cells 

 divide radially. 



In the part destined to become the capsule, the peripheral cells 

 divide at first radially, then periclinally, then radially again, then 

 again periclinally, and finally radiaUy. All these divisions occurring 

 contemporaneously m cells of the same age, it follows that the 

 outer boundmg layer of the young capsule consists of sixty-four 

 cells in cross section (that of the apophysis is formed almost always 

 of thirty-two cells). While the process sketched above is going 

 forwards, and the whole capsule is increasing in length, transverse 

 walls appear principally in the (reckoning from outside) fourth 

 layer of cells. The cells of this layer, which immediately border 

 on the central square, are originally eight, a number which becomes 

 doubled by each dividing radially, after which each daughter -cell 

 divides periclinally. The two layers so formed constitute the outer 

 spore-sac. At this time a periclinal wall makes its appearance in 

 each cell of the central square, and the cells of the third layer 

 from outside swell considerably, and, together with those of the 

 two outermost layers, grow more quickly than before in all du'ec- 

 tions. The consequence of this is that the three outer layers 

 become separated from the intrajacent ones, and this is the origin 

 of the aii--chamber. In the central square, the external layer of 

 cells is the layer of the mother- cells of the spores (spore-layer) ; 

 this forms the outer boundary of the columella — at this stage con- 

 sisting of four cell-rows in transverse section. 



In the development of the capsule, then, we see two processes. 

 The one consisting in the differentiation, from a central square of 

 four cells, of columella and spore-layer ; to this portion of the 

 capsule the author gives the term ' endothecium.' The remainder 

 of the capsule, less difierentiated but necessarily larger, comprising 

 the capsule-wall of three cell-layers and the outer spore-sac of 

 two layers with the intervening hollow space, the term * amphi- 

 thecium ' is applied. The author also observes that this latter 

 corresponds to the amphithecium of the higher Liver Mosses. 



In the layers bounding externally the air-chamber of the young 



