346 



SCIENCE 



[N. S. Vol. XXV. No. 635 



TTnless Duerden can prove that all of these 

 specimens are in some way defective and can 

 shove that the base of the third pair of ' so- 

 called primary' septa (my first pair of sec- 

 ondary septa) were destroyed, so that in all 

 cases they are shorter than the other (typical) 

 primary septa (protosepta) he can not destroy 

 the importance of this evidence. Those speci- 

 mens begin as tetrameral corals, and continue 

 so. Unless Duerden can satisfactorily show 

 that this tetrameral character of the youngest 

 stage is in these specimens a result of defect- 

 ive silicification his argument based on sec- 

 tions is incomplete. Duerden must show that 

 it is possible for a hexameral coral tip to be 

 changed by silicification, or otherwise, to a 

 tetrameral one in more than one example. 



I am not ready to concede that mere surface 

 views are unreliable. I hold that the inside 

 surface view of a young form is more reliable 

 than sections made through the tips of adult 

 individuals. In the absence of direct proof 

 to the point and yet without intending to beg 

 the question I would like to inquire how one 

 can feel sure of one's section even in forms 

 which are not accelerated? How much more 

 uncertainty must the probability of accelera- 

 tion addl 



The fact that the so-called ' primary ' septa 

 (protosepta, Duerden) under dispute have 

 positions both in locus and in sequence that 

 coincide with those called for by Kunth's 

 law, and that both of these facts point to their 

 membership among the secondary septa (meta- 

 septa, Duerden), must again be urged, and the 

 fact that the quadrant they occupy shows evi- 

 dence of acceleration must be emphasized. 

 This acceleration is shown in the case of 

 Sireptelasma rectum by the fact that these 

 quadrants generally contain more secondary 

 septa than the other two and by the further 

 fact that tertiary septa (exosepta, Duerden) 

 appear next to the counter septum before they 

 appear in the other interspaces. 



It must not be forgotten that Sireptelasma 

 profundum is an early form and devoid of 

 specialized characters and that the forms 

 studied were young ones; also that the forms 

 studied by Duerden were not only specialized, 

 but late in time (Devonic and Carbonic). 



The reassertion that the order of develop- 

 ment is that usually ascribed to zaphrentoid 

 corals does not afiect the explanation of the 

 alar fossulse as instances of retardation, nor 

 does it affect the explanation of the cardinal 

 fossula. The explanation of the cardinal 

 fossula as caused by the siphonoglyphe (sulcar 

 or sulcular) is certainly ingenious, but the 

 assumption that the gonidial groove was 

 present was made on the previous assumption 

 that the Tetracoralla are related to the 

 Zoanthese, which in turn rests upon the doubt- 

 ful ground of their primary hexamerism, 

 which was assumed from a study of sections 

 only. 



What is to be offered as the explanation of 

 the fossula of the counter septum region? 

 And if it is to be that offered for the cardinal 

 fossula — the presence of a gonidial groove — 

 what is to be advanced in support of the 

 zoanthidian relationship, since the chief 

 structural peculiarity of the Zoanthese is one 

 gonidial groove? 



When the correction of cardinal and 

 counter septum is made the cardinal septum 

 is seen to be the small one and this fact 

 would apparently support Duerden's view that 

 there was a gonidial groove in this region; 

 but I have pointed out the necessity of find- 

 ing another explanation for the fossula of 

 the counter septum or else the alternative of 

 acknowledging that both fossulse are caused 

 by siphonoglyphes, which would remove thess 

 forms yet further from their hypothetical 

 zoanthidian relatives. 



To my mind the inversion of the figures 

 counts for very little. In Duerden's criti- 

 cism at this point we find nothing that affects 

 the argument — on either side, in fact. The 

 dorse-ventral orientation is merely arbitrary. 



Such unsatisfactory definitions for ventral 

 are given for designating this aspect, that for 

 determining which to call ventral and which 

 dorsal we must decide whether, if there be two 

 siphonoglyphes, one of these morphological 

 structures is more pronounced than the other, 

 or if there be only one, whether this, in colo- 

 nial types, faces the proximal end of the 

 colony. 



It must be confessed that, so far as using 



