352 THE VOYAGE OF H.M.S. CHALLENGER. 



and tentacles very great ; usually ten to twenty very large tentacles are intermingled with 

 forty to sixty or more small ones. The size, too, of the pneumatophore is in this genus 

 larger than in all other Physalidaa and in all Siphonophorse in general; its length attains 

 in the largest specimens, fully expanded, 20 to 30 cm. or even more, its greatest breadth 

 8 to 10 cm.; the largest tentacles, fully expanded, attain a length of 20 to 30 metres 

 or even more (40 to 50 feet, L. Agassiz, 36, p. 336). 



Two species only of Caravella may be distinguished in our present incomplete know- 

 ledge of this genus, both inhabiting the Atlantic Ocean. The smaller species is Caravella 

 gigantea ( = Physalia gigantea, Bory, Physalia cystisoma, Lesson, partim) ; it occurs 

 sometimes in the Southern Atlantic, and seems to be transported from there occa- 

 sionally by westerly winds into the Indian Ocean. Captain Rabbe collected it between 

 Madagascar and the Cape of Good Hope. All parts of the body are smaller and more 

 delicate than in the following species. The number of large main tentacles is four to 

 eight, rarely more. But the principal difference is found in the isolated position of the 

 basal cormidium, which is placed at the distal end of the trunk. It is isolated by a wide 

 interval from the voluminous main mass of clustered cormidia, which occupies the smaller 

 posterior half of the ventral side of the trunk. The predominant colour seems to be 

 in Caravella gigantea more purple and violet, in Caravella maxima more blue and 

 greenish ; it is subject, however, to many variations. 



The larger species, and the largest of all the Physalidse, is the well-known Cara- 

 vella maxima, which inhabits in great numbers the Tropical and Subtropical Atlantic, 

 and especially the Gulf Stream. I observed it in great numbers during my residence in 

 the Canary Islands, in December 1866 and January 1867, and also returning from there, in 

 the Straits of Gibraltar, particularly in the Bay of Algesiras (in March 1867). It is also 

 occasionally driven by westerly winds into the Mediterranean, which it seems not to 

 inhabit permanently ; several specimens are observed in single years on the shores of 

 Italy (Naples, Messina, &c). Caravella maxima has been described as Medusa caravella 

 by O. F. Miiller and Gmelin, as Physalia caravella by Eschseholtz (1, p. 160, Taf. xiv. 

 fig. 1). Numerous later authors have figured it under the name Physalia arethusa, as 

 for example Tilesius, Chamisso, Olfers (79, Taf. i., ii.), and L. Agassiz (36, pi. xxxv.). 

 Lamarck and Lesson (3, pi. xi.) call it Physalia pelagica. (For the synonymy and 

 history of this celebrated species, compare Olfers and Lesson, loc. cit.) Caravella 

 maxima is easily distinguished from the allied Caravella gigantea by the union of the 

 basal cormidium with the other cormidia, all forming together a single clustered mass of 

 crowded appendages, which covers the greater (posterior) half of the ventral side down 

 to its basal apex. On its motions and habits, compare Olfers (79) and L. Agassiz (36 ; 

 p. 336). The young larvae of Caravella maxima, which I observed at Christmas 

 1866 in Lanzerote, are very similar to those figured in PI. XXVI. figs. 1, 2 (Cystonula). 



