THE KEAEING OF SEA-FISH LARVAE. 91 



my series. The identification of Holt's larva* with Blcnnius ocellaris 

 was admittedly tentative ; but in view of the data now available, I am 

 inclined to think that this identification should be abandoned, and that 

 the larva should be provisionally referred to B. triglo'ides. 



In support of this suggestion I would point principally to the differ- 

 ence between my specimens and Holt's description and figure in regard 

 to the position and colouration of the eye, the pigmentation of the 

 head, and the possession of preorbital and supraorbital tentacles. 



The eye in Holt's figure is set much further back than in any of my 

 larvffi ; there is no trace in the latter of the radial arrangement of facial 

 pigment which is so prominent a feature in my colleague's specimen ; 

 and the preorbital tentacle figured by Holt has no counterpart in my 

 larvce. This tentacle is not referred to in the text of Holt's description, 

 but can only represent the tentacle which in several species of Blenny 

 is situated at the anterior orifice of the nares. According to Moreau 

 the nasal tentacle is found both in B. inccqualis and B. triglo'ides, but is 

 sometimes wanting in B. ocellaris. On the other hand, a supraorbital 

 tentacle is found in B. incvqualis and B. ocellaris, but not in B. triglotdes. 

 In Holt's larva the supraorbital tentacle appears to have been wanting, 

 but is distinctly developed in my larvte, even at a smaller size. Conse- 

 quently it is probable that Holt's larva should be referred to B. triglo'ides. 



In general features, however, Holt's figure gives a fairly good idea of 

 the proportions of the body in larvre of B. ocellaris at the same stage.f 



Experiments J, K, and L. 



These experiments were begun with a new batch of larvae on August 

 23rd, after I had already gained considerable experience as to their 

 habits, in order to determine whether agitation of the water was 

 necessary or not for their successful development during the period 

 immediately after hatching, care being taken that the water should be 

 thoroughly oxygenated, and that ample supplies of food should be 

 present in the water from the beginning. 



Shallow rectangular glass aquaria were chosen for the experiment, 

 since they presented a large surface for aeration of the water. They 

 were immersed in the circulating water of the aquarium tanks, in order 

 to preserve a fairly constant temperature; and were filled with the 

 same water as was proving so successful with experiment F. No 

 plunging apparatus was set up, and no convection currents were possible 

 under the conditions of the experiment, the water being absolutely stag- 

 nant. Ten active larvt^e hatched the same day were placed in each jar. 



* I have since had an opportunity of examining this specinieu, which is far more 

 slender and of less robust appearance than my larv;c at the same stage. It should 

 certainly be assigned to a dill'erent species. 



t My friend and pupil, Mr. Arthur Darbishire, of Oxford, is kindly preparing some 

 figures illustrating the development of B. ocellaris with a view to publication. 



