324 58 



filament, scarcely reaching lo llie middle of its length {V\. Ill, fig. 11 ax). When 

 the filament is laid under the microscope the bony axis appears Iransjiarent and 

 réfringent, but its bony nature is easily shown through staining, e. g. with aliz- 

 arine or hæmatoxyline; I suppose it is the structure about which Günther says 

 (16b, p. 151): "A slightly undulated canal runs along the interior of the filament". 

 In the filaments from the middle line of the abdominal skin I have found no 

 bony axis. All the filaments are richly beset with shorter or longer branchlets 

 or twigs. Each branchlet is expanded at its end into a regular concave disc, 

 looking like a sucker; this I take to be the special organ of retention for the eggs 

 and young. I am not able to elucidate the .special manner in which the retention 

 is brought about, whether each egg is always held by one sucker, and the young 

 also, or if in the latter case two or more discs are at work. My only specimen 

 with a few eggs and newly hatched young still contained in the sac (a S. paradoxus) 

 is not well enough preserved to show things definitely; it seems to have been 

 somewhat dried and shrivelled before being preserved in alcohol, but the presence 

 of traces of discs fastened to the eggs, one on each, and also to the embryos, I have 

 ascertained with certainty. Whether the filaments are also organs of nutrition I am 

 not in a position to decide; they are pi-ovided with blood-vessels, from the stem 

 entering every branchlet; thus the nutritive function seems to me to be at least 

 possible, a secretion to the interior of the sac being probable. Certainly many 

 questions of great interest regarding the biology and development of these curious 

 fishes are to be solved, and it is to be hoped that some day one of the zoologists 

 having the opportunity of observing the living animals in their natural surround- 

 ings will take up the task. 



Possibly the whole interior lining of the brood-sac belongs to the ventrals; 

 the real condition might perhaps be, that both ventrals as in the Gobies are 

 coalesced along their upper or inner margins and the coalesced part again fused 

 with the abdominal skin; if this interpretation should prove to be the correct one, 

 the power of sending out filaments would be possessed only by the fin; this question 

 — perhaps of no great importance — could most probably be solved by an examina- 

 tion of the histology of appropriate material or of developmental stages. 



On the "thorax" only the dorsal part of the muscles is fully developed, most 

 of the lateral body-wall below the vertebral column being devoid of muscles; the 

 same is the case on the posterior, slender portion of the body with part of the 

 ventral body-wall. The myomeres are here well developed along the whole part 

 provided with dermal ossifications and further along a narrow strip close to the 

 ventral middle line of the belly, from the ventrals to about the anus; the inter- 

 vening lateral space of the body-wall, covering the side of the intestine, is — like 

 the "linea alba" — devoid of muscles. No division into myomeres is seen in the 

 strong dorsal muscles reaching from the skull to the level of the 3rd transverse 

 row of scutes; and this part of the musculature is provided with a flat strong 

 ossification (PI. Ill, fig. 10 td), corresponding to that found in Fistularia, Aulostoma, 



