59 325 



and Xerophis ophidian; narrow in I'lonl, where it firmly joins the epiolic, it widens 

 gradually backwards, finally dissolving into a number of threads. Between the 

 skull and the first transverse row of scutes it may be seen through the skin, the 

 remaining part being more or less concealed by the overlying upper members of 

 the 3 anterior transverse rows of scutes. 



The branchial cavities of the left and right sides communicate through a 

 large, oval opening (PI. VII, fig. 3 o) under the branchial skeleton, a condition only 

 found in a few other fishes (e. g. Zeugoptenis and some other flounders). There 

 are 4 complete gills (i. e. composed each of 2 rows of branchial lamellæ) (PI. Ill, 

 fig. 11 I — IV) and a well-developed pseudobranch (fig. 11 ps); the 5 gill slits on 

 each side are small, oblique, somewhat like button-holes in shape (as in Syngna- 

 thids), and as stated above not provided with gill-rakeis. The branchial laminæ 

 are rather short and clumsy, intermediate in shape between the type found in 

 Syngnathids and that of fishes generally; their number in each row from 6 to 12, 

 the anterior row of the first gill having 9, the postei-ior of the last gill only 6. 

 The pseudobranch consists of a single row of 8 leaves arranged along the inner, 

 posterior border of the hyomandibular and preoperculum, in front of the first slit. 

 As in Syngnathids the gills are not restricted to the gill-arches but for a great 

 part take their origin above and below these from the soft walls of the pharynx; 

 the rows are almost vertical in spite of the oblique position of the gill-arches."' 



As in Syngnathids the alimentary canal (PI. Ill, fig. Ill is simple, straight, 

 without any externally visible demarcation between stomach and intestine, and 

 without mesentery. The muscular oesophagus (oe) widens evenh' into a thin-walled, 

 somewhat spindle-shaped dilatation, distended with food; it suddenly narrows 

 where the intestine passes from the "thorax" into the slender part of the body, 

 just above the root of the ventrals and below the most expanded part of the air- 

 bladder. Immediately in front of this narrowing the dilatation ventrally fills out 

 the space between the two halves of the pelvis and the posterior end of the liver, 

 forming here a kind of cuneiform sacculation, no doubt only an accidental adapta- 

 tion to the given space, due to the state of stuffing with food in the specimen 

 dissected. On the ventral side of the anterior half of the spindle-shaped dilatation, 

 in the median line, the bile-duct enters (at bd in the figure), as in Syngnathids 

 the only indication of the beginning of the intestine (duodenum); thus the greater 

 part of what at first sight seems to be a "stomachal" dilatation is really formed 

 by the small intestine. The remaining part of the latter fills most of the body 

 cavity in the slender, posterior part of the body; at the spot * in the figure a 

 circular constriction marks otT the beginning of the rectum (r) *. The fold and 

 the small sacculation apparent at ** in the figure is — as far as I can see — only 

 an accidental formation due to an accumulation of the contents. 



The rectum and adjacent parts unfortunately were somewhat damaged during mj- dissection of 

 the only specimen which I could sacrifice; hence 1 am not quite sure about a few features, e. g. the 

 possible existence of an urinary Ijladder. 



42' 



