Mor^ilivhiijij of the M(t(/rejiorarii(. 239- 



arrangement i.s sucli as to leave no uncertainty that the 

 ])rimary condition was lioxaineral, and that the later septa 

 have been added in the same manner as in other forms where 

 the sequence is determinaljle from the beginning. h;uch are 

 Strepteliumii proftindinn, iStrejitelasnia loayensis^ and JJaii- 

 canella iorealis (iigs. 17-21). 



It seems unnecessary to multiply examples. Wherever 

 the tip is sufficiently Avell preserved to display the primary 

 septa they are found to be six in number ; in no case has 

 even a suggestion of a primary tetrameral condition been 

 encountered. With all these definite facts available there 

 ^vould seem to be i\o longer a possibility of any reasonable 

 doubt as to the primary hexamerism of the llugosa. 



In l)is paper Gordon does not attempt to dispute the 

 liexamerism of Lojihojihijllum prolifevum, but endeavours to 

 explain it as the result of the precocious appearance of what, 

 according to him, should be the first pair of metasej)ta. 

 'J'liis pair, here regarded as the primary dorso-Iateral pair, 

 Gordon supposes to belong not to the protosej)tal, but to the 

 metaseptal series, and owing to its accelerated appearance it 

 gives a false hexameral character to the primary stage. 

 This idea of acceleration is altogether hypothetical, and its 

 author does not produce a single acceptable fact in its 

 sup{)ort. He considers that a departure from the original 

 tetrameral type is likely to occur in a form such as LopJiO' 

 phylhun which appears in Carboniferous times, that is, towards 

 the close of the geological distribution of the rugosids. This 

 argument, unsatisfactory in itself, now fails altogether in view 

 of the fact that comprised in the list of corals given above, 

 in which six primary septa have been defii\itely established, 

 there are representatives of almost all ages in the chrono- 

 logical extension of the rugose corals. 



Gordon draws attention to the fact that in the older stages 

 of growth of a corallum there is no diiference in character 

 between the dorso-lateral pair of primary septa and the 

 principal septa which arise later, that the niterseptal spaces 

 between these septa and. the adjacent septa differ in no ways 

 from the other interspaces, and that exosepta appear in the 

 dorso-lateral primary interseptal spaces just as in others. 

 Were these really primary septa he expects that they would 

 present some feature distinguishing them from the later 

 principal septa. Such an expectancy, however, is altogether 

 contrary to what we actually know of development and growth 

 in the Anthozoa. All studies in this group, particularly among 

 the aclinians and corals, reveal that the growth tendency is 

 everywhere towards a perfectly cyclic plan, with all the parts 



