312 PEOF. p. M. BTJNCAN ON THE 



to meet. The synapticula are in close and regular succession m 

 some places, are nearly as high as the septa are broad, are of 

 course narrow on account of the proximity of the septa, and 

 are never continuous, as in Fungia. Usually the synapticula 

 are slightly dice-box shaped ; and there may be from eight to 

 twelve in a vertical series in the interseptal locu.lus on each side 

 of the axis of the colline. 



The coUines form more or less convex projections ; and as 

 they consist of septa placed side by side and very close, it is evi- 

 dent that the synapticula nearly fill up the interseptal loculi. 

 Is there a wall separating one side of the colline from the 

 other, and reaching along its long axis from the basal lamina to 

 near the free surface ? 



There is no true wall, but a long synapticnlum more or less 

 discontinuous is in the vertical line, and it is tolerably broad. 

 The synapticula already noticed are on either side of it near the 

 junction of the septal margins (vertical) with the columella. 



The columellary structure was clearly indicated by Dana in a 

 small drawing * of the structures of Agaricia riigosa, which is 

 probably Pachyseris Valenciennesi. In the specimen now under 

 consideration the columella is a mere groove above ; but a sec- 

 tion across the line of the collines and valleys shows that the 

 columellary space is crossed by numerous tabulate-leoking pro- 

 cesses which stretch from the free edge of one septum to another. 

 In number these processes tally almost invariably with the 

 synapticula in vertical series ; and as these last are generally on 

 the same level in the interseptal loculi, it happens that the colu- 

 mellary synapticula start from the long inner surface of the level 

 synapticula and corresponding free edges of the septa. So that 

 long tabulae are produced one over the other, and closing in the 

 narrow axial space. The uppermost forms the visible columella. 



The granulations on the sides of the septa are exceedingly 

 developed in young or outer series ; but it is perfectly evident 

 that they form no part of the synapticular structures. The septa 

 are rarely perforate. 



The basal lamina is very solid, and its growth aj)pears to be 

 truly thecal, and not synapticular. 



There are no endothecal dissepiments in PacJiyseris ; and the 

 interseptal loculi, instead of being open down to the basal wall, 

 are partly occluded by vertical series of synapticula; and the 



* Dana, op. cit. pi. 22. fig. 1 b. 



