366 DISCOVERY REPORTS 



and a " basal cap " (Fig. 29, phot.c.b.c.) at the end of the cell remote from the nucleus. 

 The photogenic cells are arranged parallel to each other and form the units of which 

 the organ is built up. Their disposition, however, differs in a constant and surprising 

 manner from that found in the transverse streak both of S. affi?ns and of S. debilis 

 itself, and in fact in all the photophores so far encountered which are composed of 

 similar cells. The photogenic cells are so disposed that their nuclei occupy the basal, and 

 not the distal, ends of the cells, which are thus completely inverted with respect to the limb. 

 This is a most astonishing and inexplicable feature, indicating strikingly the need for 

 great caution in any speculation regarding the optical properties of any particular pattern 

 of photogenic cells. 



No granular deposits such as have been observed in other organs have been dis- 

 covered at the ends of the cells occupied by the basal caps, or indeed anywhere else in 

 the organ. Investing the whole photophore is a loose fibrous layer, most strongly 

 developed at the superficial ends of the photogenic cells. Among the fibres of this 

 layer numerous compact nuclei lie behind and at the sides of the photophore, but no 

 nuclei have been observed between the photogenic cells and the integument, indicating 

 the probable origin of the organ as a specialized portion of the chitogenous epithelium. 

 The connective tissue layer is no doubt pigmented during life and acts as a reflector 

 or at any rate as a screen. 



No nerves have been found in association with this photophore, but this failure may 

 be perhaps due to the somewhat indifferent fixation of these specimens. There is no 

 thickening of the integument in the form of a lens above the organ. 



(d) The carapace photophores 



The carapace photophores of S. debilis are disposed near the lower margin of the 

 branchiostegite (Kemp, 1906) in a pattern much like that formed by the organs of 

 S. affinis (Caiman, 1939), and from this it might be expected that they would possess the 

 same structural peculiarities. Actually they are basically similar to the carapace organs 

 of Hoplophorus, differing only in their greater simplicity. 



They consist of a number of photogenic cells arranged in a single linear series, the 

 number of cells in a row varying from three to twenty-nine in the specimen I examined. 



The nuclei of the photogenic cells (Fig. 30 A, B, n.phot.c.) are finely reticulate and 

 are smoothly ovoid in shape, thus differing markedly from those of the carapace organs 

 of Hoplophorus, which possess concave proximal surfaces. No trace of cytoplasm or 

 contents other than the nucleus have been observed in the cells, which therefore show 

 the clear area (Fig. 30 A, B, phot.c.cl.a.) characteristic of photogenic cells of this type. 

 No "basal cap" nor any trace of nerve fibres have been discovered in connexion with 

 the cells, which are clearly much simpler than those mentioned in Hoplophorus. 



The integument is slightly raised by the underlying photogenic tissue (Fig. 30 A), 

 but there is no marked thickening or differentiation of the chitin in the form of a lens. 

 Beneath the integument forming the outer surface of the branchiostegite and that lining 

 the branchial chamber lie normal nuclei of the chitogenous epithelium (Fig. 30 A, B, 



