20 BULLETIN: MUSEUM OF COMPAEATIVE ZOOLOGY. 



the median portion of it is an extension of the parasphenoid. I know of no other 

 instance where the parasphenoid extends anterior to the vomer. When the pala- 

 tines have lost their connection with the pterygoid it seems possible that they 

 may have swung inwards at their inner ends and become attached to the paras- 

 phenoid. This theory is suggested only as a possibility. It seems more rea- 

 sonable to suppose that the palatines have failed to ossify, and that the above 

 process is of the parasphenoid. 



The lateral head bones are bent forward and downward against the large 

 mandible, and all of them are anterior to the posterior end of the mandible. The 

 mesopterygoid and metapterygoid are absent. 



The usual opercular bones are present, but so bent forward that the preopercle 

 is inferior to the others instead of in front of them as is usual. The ligament 

 connecting the interopercle with the angular has ossified for the greater part of 

 its length. 



The hyomandibular and quadrate are elongate bones, but normal in their rela- 

 tion to each other and the other bones. The former has a simple rounded head 

 for articulation with the cranium, no mesopterygoid process, but with the usual 

 processes for the opercle and preopercle, though these are only slightly devel- 

 oped. The symplectic is slender and not much ossified ; it extends downward 

 behind the quadrate. The pterygoid is a very small sliver of bone closely attached 

 against the anterior edge of the quadrate. 



The three usual bones make up the enormously developed mandible. The 

 dentary is abruptly narrowed at its anterior end to form the small mouth. 



The premaxillary is a broad fan-shaped bone with a long process extending back 

 at the side of the ethmoid to the top of the head. The maxillary has a similar proc- 

 ess extending upward behind and parallel with the premaxillary process. The 

 opening of the mouth apparently depends to a large degree upon the maxillary 

 elements, as the mandible is connected by skin rather closely to the side of the head. 



The condition existing in the hyoid apparatus is unique. Each side of the hyoid 

 arch lies wholly in front of the branchial arches, its lower edge on a level with the 

 greatly enlarged first basibranchial (glossohyal). It is unattached and widely 

 separated in front from its opposite fellow. As the branchiostegal rays are on the 

 upper edge of the cerato hyal and inclined upward rather than on the lower edge 

 and inclined downward, as is usual in other Teleosts, the movement of the hyoid 

 must be outward from the upper edge instead of from the lower, thus enabling 

 the fish to spread the lateral bones of the head wide apart somewhat as is shown 

 in the picture published by Shaw (Trans. Linn. Soc Loud., 1791, vol. 1, pi. 6) and 

 copied by Goode and Bean (Oceanic Ichthyology. Mem. Mus. Comp. Zool., 1896, 

 vol. 22, pi. 116, fig. 394). The interhyal is a very long fan-shaped bone with its 

 upper end uneven and fibrous and loosely attached so that this outward movement 

 is not retarded. 



The first two branchial arches are attached to one large basibranchial. Anterior 

 to this is a very long pen-shaped bone, nearly as long as the cranium. It is prob- 

 ably the homologue of the glossohyal, though it is unattached to the hyoid arch. 

 The hypobranchial of the first arch only is present. All of the superior pharyn- 



