1740 THE VOYAGE OF H.M.S. CHALLENGER. 



and is closely allied to the preceding Coelodendrida ; they differ from the latter mainly 

 in the development of a peculiar new organ, the " rhinocanna," or " nasal tube." This is 

 a hollow tube placed in the sagittal plane, arising from the base of each galea, and is con- 

 nected with its apex by a simple or double frenulum. Between the oral openings of the 

 two opposed rhinocannse (one dorsal and one ventral) lies the proboscis of the central 

 capsule. 



The first observed species of Ccelographida was Coelographis gracillima, some parts of 

 which (but not the entire skeleton) were figured in my Monograph (1862, Taf xxxii. 

 figs. 2, 3). But I confounded these with Ccelodendrum gracillimum, in the branched 

 hollow trees of which the fragments of the former were entangled. I detected this error 

 afterwards, when I had the opportunity of observing some complete specimens. The first 

 description of a complete skeleton was given in 1882 by 0. Biitschli, who examined a 

 large specimen of Ccelothamnus davidoffii, cai:)tured by Davidoff in the Mediterranean 

 (Zeitschr. £ wiss. Zool., Bd. xxxvi. p. 486, Taf. xxxi.). In the rich collection of the 

 Challenger I was able to distinguish not less than nine genera and twenty-six species of 

 Ccelographida, but the majority of their large and most fragile skeletons were more or 

 less injured, or quite broken. It was, therefore, of the highest importance for the 

 minute study of this diflicult group, that Dr. John Murray, during his expedition to the 

 Fseroe Channel (in 1882, in H.M.S. "Triton"), discovered in the Gulf Stream the 

 beautiful Coeloplegma mun'ayanuni, and brought up home hundreds of well-preserved 

 specimens (PL 127). Only by the complete examination of this excellent material it 

 was possible to answer manj^ diflicult questions as to their morphology, and to correct 

 the errors in my description and in that of Biitschli. 



We divide the family Ccelographida into two rather different subfamilies, which 

 may afterwards be separated as two divergent families, the Coelotholida (PL 122) and 

 Coeloplegmida (Pis. 126-128). Both groujjs may be easily distinguished at a glance, 

 since the numerous branches, arising from the hollow radial tubes, remain constantly 

 free and independent in the former, and represent a spiny thicket, whilst in the 

 latter they constantly become united, and by anastomosing form a peculiar " mantle," 

 or outer envelope of delicate network. But besides, there are other and more important 

 differences between the two groups. The peculiar hollow tube, arising from the base of 

 the galea on each valve, which is filled with phseodella, and which we call the rhinocanna, 

 develops in the Coelotholida on its open mouth two paired lateral frenula (right and 

 left), which connect it (like two lateral bridges) with the base of two paired hollow 

 main tubes (the "frontal tubes"). In the Coelospathida, however, the mouth of the 

 rhinocanna develops a single odd frenulum only (in the sagittal plane), and is connected 

 b)^ it with the base of an odd, single, hollow main tube, dii'ected forwards, the " nasal 

 style." 



The central bivalve lattice-shell, from which the galea and the tubes arise, exhibits in 



