496 ME. H. M. BERNAKB ON 



open, or else, especially in cases where the horizontal elements 

 are flaky, it may appear solid. I have never seen the columellar 

 tangle as such protuberant or convex in Porites. 



And here it is worth noting that there appears to be some 

 trace o£ dimorphism in the ealicles of a stock, while in some the 

 centre of the columellar tangle fills up with cross bars, in others 

 the latter remains a deep open pit. These two forms on one and 

 the same stock are so frequent, that they suggest some definite 

 physiological significance {cf. infra). 



The tubercle which rises from the columellar tangle but is 

 absent when the centre is hollow, is, as above noted, often found 

 flattened in the directive plane and here and there joined to one 

 of the directive pali. It is the equivalent of the directive keel 

 seen bisecting the columellar tangles in many Turhinaria. It 

 appears as if, in the earliest stages of budding, the directive septa 

 met straight across the calicle, and that this columellar tubercle 

 is the remains of the connecting link. This conclusion commends 

 itself from the fact that a close study of columellar tangles, aa 

 in Turhinaria, where they are specially well developed, shows 

 that they are primitively built up of the usually curled lower 

 edges of the septa. 



These comparisons help to confirm my view of the relationship 

 of Forites to the Madreporids. In the case of Madrejpora, the 

 directive plane in the buds falls in with the radial symmetry of 

 the parent, and might almost be regarded as coincident with that 

 of one of its costse. In Turhinaria although, in large stocks, the 

 directive keels point almost all ways, yet, near the margins, 

 they are often seen all pointing towards the growing edge. I 

 have been fortunate enough to see this phenomenon in a thin 

 explanate Porites (fig. 4, PI. 35). In this figure the great 

 majority of the directives point nearly uniformly up and down, 

 that is, in the line of growth. If we refer to the diagrams, the 

 direction of growth runs out between the four principal pali as 

 indicated by the arrow (diag. 4 & 5, p. 492). 



Whether these directive planes, when found twisted all ways 

 on a stock, indicate special relationships between adjacent indi- 

 viduals of the colony it is impossible to say. I should be inclined 

 to think that the individual as such is submerged except in 

 Madrepora, and that the buds come from the common coenosarc. 



Before leaving this columellar tubercle, I w ould draw attention 

 to the fact that I have inserted it in all the diagrams as a per- 



