1884.1 5<J& [Cope. 



two cephalic arches, both freely movable ; (1) an anterior dentigerous one — 

 the palatine, and (2) the suspensorial, consisting of the hyomandibular 

 and quadrate bones ; without maxillary bones or distinct posterior bony 

 elements to the mandible ; with an imperfect scapular arch remote from 

 the skull ; and with separately ossified but imperfect vertebrae. " 



M. Vaillant came to no conclusion as to the affinities of this group ; and 

 Messrs. Gill and Ryder remark, "We are unable to appreciate any affinity 

 of Gastrostomus to any Anacanthines, Physostomes, or typical Apods, 

 nor does it seem to be at all related to Malacosteus, which has been 

 universally considered to be a little modified Stomiatid." It is, how- 

 ever, clear to me that the relationships of this family Eurypharyngidse 

 are to the order Colocephali, and that they represent the extreme de- 

 gree of the modification of structure which that order exhibits. In 

 other words, the modification of the ordinary piscine type which is 

 iound in the Anguillidae (order Enchelycephali), is carried to a higher 

 degree in the Colocephali, and reaches its extreme in the Eury- 

 pharyngidse. The points of identity between the two groups last-named 

 are so many, that it becomes desirable to ascertain whether they are 

 susceptible of ordinal separation from each other. The characters 

 above given to the order Lyomeri are in fact identical with those which 

 define the order Colocephali, with a few possible exceptions. Eirst, how- 

 ever, I note that the supposed palatine arch, is probably the maxillary, 

 as in the Colocephali, and that it is the palatopterygoid arch which is 

 absent. The five branchial arches exist in the Colocephali, but the three 

 anterior are rudimental, and the basal branchihyal bones of the fourth 

 and fifth are closely united. There are, however, five arches. There is a 

 ceratohyal arch in Mursena and Gymnomursena, but of very slender pro- 

 portions. Whether this element is absolutely wanting in Gastrostomus, 

 or whether the first branchial arch is its homologue, remains to be ascer- 

 tained. Should the last two be coherent as in the Colocephali, we would 

 then have the same number of hyoid arches in both, viz., six. The "im- 

 perfectly ossified cranium " is shown in the detailed description given by 

 Messrs. Gill and Ryder, to support the same bones which are found in the 

 Mureenoid skull. The degree of ossification of the skeleton does not con- 

 stitute a basis for ordinal distinction, if the same elements be present. 

 For this reason the perforation of the vertebral centra by the remnant of 

 the chordadorsalis does not seem to be of ordinal importance. 



In the more detailed description, there are a few charecters worthy of 

 notice. First, "The notochord is persistent in the skull for half the 

 length of the basioccipital." This indicates further the primitive condi- 

 tion of the vertebral column, but scarcely gives basis for an ordinal defi- 

 nition. Second (p. 266.), "The neurapophyses are slender, diverging 

 (instead of convergent), cartilaginous distally, and embracing the neural 

 sheaths on the sides, while by the neurapophyses is supported a membra- 

 nous sheath which roofs over the nervous cord," etc. The nerual canal 

 is well closed above in the Muramidse, but in the Anguillidae it is largely 



