98 MADREPORARIA. 



due to the rolling of the specimen during life, the uppermost point acquiring a cap of fresh 

 growtli. 



The calicles vary from I'f) to 1-75 mrn., probably not very deep, with very thin ridged 

 walls, often showing an irregular zigzag. The septa also very thin and perforated. The typical 

 formula with the six pall, and the directives can be made out. Columella composed of little 

 more than the pali and the central tubercle, and the few threads which connect them at intervals. 



This diagnosis has been arrived at only with difficulty. The exposed surfsices, which 

 appear to be well preserved, are quite misleading. They show the walls as thick, round- 

 topped ridges, with the elements of the skeletal reticulum also tliick. They had, however, 

 clearly suffered from corrosion before fossilisation (see above, p. 22). Tlie sections, 

 again, near the surface have been confused by infiltration. The true character of 

 the original skeleton is, however, fortunately, discoverable on a section passing some 

 4-5 mm. below tlie surface. The figure shows indications of the fusing septa (which by study 

 of many calicles can be seen to have been typical), the pali, and the central tubercle. These 

 characters are enough to establish the right of the coral to a place in this genus. More details 

 than those shown can be made out by comparing calicle with calicle, and the zigzag thread 

 of the wall can also be established ; although it is not well shown in the figure, where the 

 calicles are in contact. 



There is, however, another specimen which I think is of the same kind, a small nodule 

 broken off from a larger stock. This has exposed calicles showing the character of the skeleton 

 as given in the diagnosis, except that the skeletal elements have been thickened. Taking this 

 fact into account, we can make out the zigzag wall forming a low sharp ridge, the perforated 

 septa, somewhat confused owing to the] post-mortem thickening of the skeletal elements, and 

 the six pali rather swollen. 



a, h. rxeol. Dept. E. 4810. 



(One sliced fragment belonging to a, with typical calicles marked.) 



73. Goniopora Persia (4)4. 

 [Guverchin Kala, N.W. of Lake Urmi (Miocene), coll. Loftus ; British Museum.] 



Description. — Corallum appears to have been an erect, rounded or oval nodule, laterally 

 compressed, some 4 cm. high, 4 cm. broad, and about 2-5 cm. thick. The vertical section 

 shows the calicles radiating from the point of attachment. 



The calicles were about 2-5 mm. across. The walls were thin, zigzag, very perforated 

 and incomplete. The septa were also very thin and perforated, and arranged in the typical 

 formula, with the full number of pali and the directive septa. Columellar tangle very slight, 

 and consisting chiefly of the pali and a central tubercle. 



I was at first inclined to regard the white pattern of the section as representing tlie coral 

 skeleton, as Dr. Abich may have done (see above G. Persia 2), but my experiences with the 

 fossils last described {G. Persia 3) have led me to see that, in this case at least, the thin dark 



