338 Prof. E. W. MacBride. The Artificial Production of 



Since, moreover, the growth of Nitzschia varies immensely with conditions, 

 too much or too little light being equally deleteiious to it, it is very difficult 

 to ensure the co-operation of this factor in the right strength at the right 

 time. 



The appearance of a single larva with two hydrocceles in the control 

 plunger jar serves to warn us that other stimuli besides increased salinity 

 may give rise to this phenomenon, and the circumstance that embryos with 

 two fully developed hydroeoeles were observed by Metschnikoff (14) among 

 the specimens which he extracted from the womb of Amphiura squamata 

 tends to emphasise this point. Indeed, when we reflect that most cases 

 recorded of larvse with two hydrocceles have been represented by the occur- 

 rence of a single specimen amongst hundreds of normal larvse, we may feel 

 certain that their development has not been caused by increased salinity. 



We may now turn our attention to the structure of the abnormal larvae 

 which were obtained. In a minority both hydrocceles were equally developed 

 (fig. 8), and in these cases the characteristic pedicellariai of the right side 

 were absent. The right hydrocoele had induced the formation of a right 

 amniotic cavity, into which its outgrowths projected as a second series of 

 primary tentacles, and from the floor of this cavity pointed spines were 

 produced similar to those belonging to the Echinus-rudiment on the left side. 

 When such larvae were explored by sections, it was found that on the right 

 side the rudiments of a second Aristotle's lantern were represented by dental 

 pockets just like those on the left side {d.s., fig. 19), and that the right 

 hydrocoele was connected with the right anterior coelom by a well formed 

 right stone-canal. The right anterior coelom had been enlarged so as to form 

 a right axial sinus which had become fused with the left axial sinus, and 

 thus there arose a large median cavity lying above the oesophagus; and the 

 madreporic vesicle, which, it will be remembered, normally lies above the 

 oesophagus, was sometimes entirely suppressed (fig. 20), but sometimes 

 persisted as a narrow cavity intervening between the compound axial sinus 

 and the gut (tigs. 18, 21). 



When describing the development of Asterina gibbosa (11), I stated that 

 the madreporic vesicle was a rudiment of the right hydrocoele, because in 

 the abnormal specimens which possessed a fully developed right hydrocoele I 

 was unable to detect a madreporic vesicle. No doubt in these specimens 

 the right hydrocoele had completely suppressed the madreporic vesicle as in 

 the larva represented in fig. 20. 



The compound axial sinus may communicate with the exterior by one 

 (fig. 20) or two pore-canals (fig. 21). But in those larvae in which I found 

 two pore-canals both of them belong to the left side, and the second pore- 



