'37 



testicular, and in two larger polyps (calicle about 23 mm. long 

 entirely formed of ova. 



I had not sufficient examples of larger sized corallites as 

 obtained in dredging No. 3, which I could decalcify so as to 

 trace the changes in generative organs with increase of size. 

 I am, however, impelled to consider that there must be pro- 

 tandry. The polyp first produces testicular elements, w^hich 

 are replaced as it grows by ova ; a regular crop of these are 

 then ripened. With increase of size the rate of growth of the 

 corallite seems to gradually lessen. This is correlated with 

 the production of ova, the increase in the number of which 

 causes cessation of growth and finally the death of the parent 

 polyp*. 



Note on F. PAVONIXUM.— I have only been able to 

 afford to use one polyp between Nos. 6 and 7 of the table of 

 measurements on p. 124 for the study of the anatomy of this 

 species. Tentacles are present over all the septa, and the 

 latter are all entocoelic, there being thus relatively twice as 

 many mesenteries as in F. rubntni. Those pairs of mesen- 

 teries, which lie on either side of the septa fusing by tra- 

 beculae in the axial fossa, alone appear as a rule to depend 

 from the stomodoeum. 



In all other respects the anatomy is the same as above 

 described in F. riibrum. The polyp is in the male condition. 

 The spermagens are tightly packed together, and present in 

 side view a round to branched appearance. 



7. MINUTE ANATOMY OF THE POLYPS OF F. 

 RUB RUM. (PI. I and II, figs. 1—9). 



Calicoblastic Ectoderm figs. 1-3). — The layer of ectoderm 

 separating the polyp from the corallum is everywhere com- 

 plete, and even in the most roughly decalcified specimens is 

 not torn away. It varies considerably in accordance as it 

 may be in any position an active secretory layer or not. No 

 definite cells can in any part be distinguished. Over the 

 greater part of the corallum it is an extremely thin, finely 

 granular layer, slightly thickened where nuclei are present. 

 The latter are generally slightly oval in shape with granules 

 but seldom a network. It only differs from the same layer 

 in other corals in being better defined and more definite. 



Near the base of the polyp and on the sides of the septa 

 the calicoblastic layer simulates the appearance of a pave- 

 ment epithelium, nuclei joined together by finely granular 



* Since the above was wiittcn I have examined a large number of specimens of 

 llie same Coral frfm other localities. Vxde "Some Notes on Variation and 

 Pretandrv in I'lahcUum rubruin and senescence in the Sjme and other Corals." 

 Troc. Camh. Phil. Soc, vol. XI , pp. 463-71 (1902). 



