SALPA. 11 



in direct communication with no less than three adjoining ones, as 

 will be seen hereafter. When seen in i)rotile (Fig. 4), the outline 

 is ellipsoidal ; the two principal openings are placed at a distance 

 from the extremities, the anterior spur of the tunic extending be- 

 yond the opening, thus bringing both their openings rather more to 

 the ventral side, and not strictly along the continuation of the axis, 

 but on each side of it. There are only five muscular liands, one at 

 the posterior extremity, three others uniting on the ventral surface, 

 somewhat behind the anterior part of the gill, and another ill-defined 

 one at the anterior extremity. The nucleus (w.) is much larger in 

 proportion to the body than in the solitary form ; the endostyle (e.) 

 occupies but a little more than a third of the cavity. The gill, when 

 seen in profile, runs somewhat obliquely towards the anterior ex- 

 tremity, where it is nearest the ventral surface. There is no per- 

 ceptible difference in the size of the anterior and posterior openings 

 of the respiratory cavity. The vibratile cavity and the nervous 

 ganglion do not differ in structure from those of the solitary form ; 

 the languet is perhaps somewhat broader and more powerful. In 

 the chains I have had the opportunity to examine, I found either 

 that the solitary foetus had already lieen expelled, or was only 

 slightly developed, so that I can only say that its position corre- 

 sponds with what has been described by Sars, Krohn, Vogt, and 

 others, the tests, as is well known, being greatly developed in the 

 individuals which had already lost their solitary foetus. What is 

 worthy of special notice in the aggregate form is the great thick- 

 ness of the tunic ; this would make the connection between the 

 individuals of a chain simply a mechanical one, were it not for the 

 spurs from the respiratory cavity (s. c), which project through the 

 thickness and connect with similar spurs in adjoining individuals. 

 The spurs disappear invariably after the individuals of a chain have 

 become separated for any length of time, and they are incapable 

 of reuniting again, as has erroneously been asserted. When thus 

 freed, the aggregate form is perfectly helpless, the great thickness 

 of the tunic preventing it from regulating its motion ; while, when 

 connected as a chain, their capacity to guide the chain in any par- 

 ticular direction is much greater. 



Sars has described exceedingly well the mode of aggregation of 

 the chain of Salpa runririnata; the chains, however, were quite 

 far advanced, and he found it impossible to trace distinctly their 

 mode of juncti(_)n. Soon after the chain escapes from the solitary 

 form, while still quite small, so that foui- or five individuals can 



