50 BULLETIN 84, UNITED STATES NATIONAL MUSEUM. 



The first under brachial plate is very small and extremely narrow, chiefly in its 

 distal part, where it is pressed between the external lobes of the adoral plates; 

 proximaUy it becomes wider and ends in a very convex margin. The following 

 plates are, as in A.jlexuosa, almost quadrangular and wider proximaUy; the proximal 

 side is also bent into three unequal sides, so that the plates finally display a hexag- 

 onal shape. They are as wide as long or a httle longer than wide; however, they 

 are proportionately a httle less elongated than in A. flexuosa. 



The lateral plates are little protruding, but they cover a rather important part 

 of the upper face of the arms; they first carry six spines but this figure afterwards 

 diminishes to five; these spines are subequal and more or less flattened. The first 

 ventral spine, the length of which somewhat exceeds that of the article, is very 

 much flattened ^vith a rounded point. The second one is a little shorter and rather 

 cylindrical; on the first articles the spine simply ends in a point, but from the 

 seventh or eighth article upward it shows a hyaline terminal hook which rapidly 

 takes on a very great development; this bent hook is analogous to the one which 

 is known in A. complanafa. However, the angle formed by the hook with the spine 

 is less marked than in this latter species. Ljungman says nothing of the hook in 

 his original description, but in the table of the species of Amphiuridffi, published 

 by him in 1871, he indicates it as follows: ''Spina brachialis ad infimam proxima 

 falciformis in apice acuta." The other spines are more flattened and they tend to 

 assume a lanceolate shape; their Avidth very slightly decreases down to the last 

 one, v/hich is short and very much widened. 



There are two tentacular scales, very small, rounded, and subequal; the proxi- 

 mal scale inserted on the lateral plate is, however, a little larger than the other; it is 

 disposed obliquely to the latter. These scales are scarcely visible, and they may even 

 be completely absent on the first three or four brachial articles, but they always 

 appear afterwards in a regular maimer and I can not account for Ljungman's 

 saying that they are lacking. 



By the above description, it will be seen that A. latispina shows great analogies 

 with A. flexuosa, but it differs clearly from it by the shape of the brachial spines 

 and by the large bent hyaline hook which terminates the second ventral spine. As 

 to the peculiar disposition of the upper plates of the disk, and the absence of these 

 plates on the iiiterradial spaces, which somewhat recall the genus Ophionephthys, it 

 would be necessary to compare several specimens in order to know if this disposition 

 be accidental in the only Icnown specimen or if it does really characterize the species. 

 The former hypothesis seems to me the most likely one. H. L. Clark pointed out 

 in A. acrystata from California and Japan the same disposition of the plates m some 

 specimens (11, p. 146, fig. 58 a), whereas in others (fig. 58 g) the plates cover the 

 whole upper face of the disk. 



AMPHIDRA RATHBUNI. new species. 



Plate 18, figs. 5 and 7. 

 Amphiura flexuosa H. L. Clark (not Lyman) Bull. U. S. Fish Comm. for 1900, vol. 2, 1901, p. 247. 



Fish Hawk station 6066. Jan. 20, 1899. Mayaguez Harbor, Porto Rico; 16 

 to 17 fathoms; m. s.; temp. 23° C. One specimen. 

 Type.—C&i. No. 21295, U.S.N.M. 



