132 



different stages of development in a tow-net attached to the 

 beam trawl in about 20 fathoms in False Bay, off the Roman 

 Kock, 13th October, 1898. 



The smallest of these closely resembled the larvae of Species 

 I., and the largest showed the widely separated ventrals with 

 adhesive apparatus between them and the short dorsal fui situated 

 on the tail, characteristic of the family Gobiesocidae. The 

 ethers represented intermediate stages. 



The only member of this family recorded from the Cape is 

 Chorisochisnuis dcntcx, but as the posterior half of the ventral 

 disc in these young forms has a free margin, they cannot be re- 

 garded as belonging to this fish, and they are not sufficiently far 

 advanced to allow of more than a reference to the family with 

 certainty. They have a gill cover free from the isthmus, and 

 therefore do not belong to a species (apparently new) of Lepado- 

 gaster recently found in False Bay. In view of the characteristic 

 oval eggs of this last-named genus,* a number of eggs were re- 

 examined. In Part I. they were described as about one milli- 

 metre in diameter. A number of measurements show that none 

 of them are perfectly circular, though some are ver}' nearly so, 

 C'ne being ro6 x "98 mm., while others varied from about this 

 to I '37 X -gj. In most of these eggs there was also one 

 oil globule from '\y to "3 mm., and the space between the eggs 

 was less than in the case of those at first examined. In the 

 fresh egg there was no evidence of a filamentous fringe round 

 the basal part of the egg capsule, though some preserved in 

 formalin showed radial striae on an irregular border. 



The identification of the other demersal egg found (Species 

 11.) has not yet been possible, but several speciniens were 

 again procured in dredging on rough ground in False Bay in the 

 month of November. That they belong to a fish of small 

 dmiensions seems probable, as they have been on more than 

 one occasion found inside an empty barnacle shell, the opening 

 of which was small. Plate V, fig. 34, represents such a shell, 

 (natural size) with one side cut away to show the blue eggs 

 attached to its inner surface. Fig. 35 represents one of the 

 eggs enlarged to show the numerous oil globules which occur 

 towards one side of the egg. 



In the following account, the various species dealt with are 

 arranged according to the number of the oil globules and 

 diameter of the eggs, as this artificial arrangement has proved 

 very convenient. For this reason also a key (p. 150) to the e.^gs 

 and larvae similar to that given in Part I., has been drawn up, 

 including not only the species here dealt with, but those already 

 mentioned in Part I. The items taken from the first table are 

 printed in italics for convenient reference. 



* See Holt, Trans. K. Dublin Soc, S. II., IV., iSyi. 



