558 BULLETIN 188, UNITED STATES NATIONAL MUSEUM 



imison, like oars in rowing a boat. They make a characteristic trail 

 on the mud : a central streak where the body drags and pairs of de- 

 pressions on either side where the pectoral fins are inserted. They 

 are somewhat clumsy on banks with even a moderate slope and often 

 roll over on their backs but quickly right themselves. 



By far tlie best account of the habits of this fish has been given by 

 Johnstone and is attributable to Annandale and Robinson ; the follow- 

 ing excerpts are from that account : 



The adults are very abundant in the immediate vicinity of the town of Jambu, 

 where a large area of semi-liquid mud is exposed at low tide, but their wariness 

 renders them difficult of approach, especially at half-tide, when they do not appear 

 to have settled down to a terrestrial existence. They seem to be vei'y sensitive 

 to cold or fog, and do not come out on the mud-flats early in the morning, even 

 at low tide. When the sun is well up, however, they skip about many yards 

 away from any water, and quite careless as to whether their tail or any other 

 part of their body can be kept wet, even when the temperature is so high that 

 sand becomes too hot for a European to walk upon it with bare feet. * * * 



A most important feature in which the habits of the present form differ from 

 those of its congeners is that it constructs a more or less permanent burrow, in 

 which it takes refuge when alarmed and remains in bad weather. Whether this 

 burrow is also used as a nest we cannot be sure, but it appears to be frequented 

 by individuals of both sexes, and the natives say that the female deposits her 

 eggs in smaller holes made for the purpose. The main entrance to the large 

 burrows is always more or less funnel-shaped and slanting, and lies in an oval 

 or circular pool of water, which is retained round it as the tide sinks — for it is 

 always made between tide-marks — by a rampart of mud four to six inches in 

 height, and about a foot and a half in diameter. It is surrounded, outside the 

 rampart, by a number of simple holes, which probably lead into the central 

 shaft, though, owing to the liquid condition of the mud, we were unable to 

 demonstrate this by digging. When thoroughly alarmed, the fish sometimes 

 enters one of these holes, but if only a little scared, it rushes to the rampart, 

 surmounts it, and sits for a longer or shorter period on it before diving into the 

 pool and disappearing. It seemed quite evident to us that each individual had 

 its proper burrow, but how many inhabited one burrow we could not ascertain , 

 it appeared that their number must be small. The central pool was always 

 entered at one point, so that a definite furrow was made on the rampart, and 

 the characteristic tracks of the fish — consisting of a central line (made by the 

 tail) and two series, running parallel to one another on either side of the line, 

 of more or less hand-shaped depressions (made by the tips of the more prom- 

 inent pelvic rays and the web between them) — converged towards it, showing 

 readily whether the fish had recently gone out or in. 



At first we were much puzzled as to how the rampart and burrow were made, 

 but the natives told us that the fish wriggled down into the mud, filled its 

 mouth with the stiffer clay beneath the surface, and built up the wall with 

 pellets thus brought from below. We foimd it impossible to confirm their 

 statement by actual observations, but there is every reason, in this case, to 

 regard them as accurate observers, for the colour of the walls plainly showed 

 that they were not made of surface mud, while the shape and size of the pellets, 

 which could easily be distinguished in freshly constructed specimens, were just 

 what would have been expected had they been casts of the inside of the mouth 



