276 SMITHSONIAN INSTITUTION 



because we have since learned that all sutures in this genus are more or less sinuous at the 

 outer surface and straight underneath, so that this character is a mere matter of preservation. 

 Angelin's figures of his specimens were not accurate as to this; both of them show waving 

 sutures in many places, as will 'be seen by comparison of the new figures now herein given 

 from the careful drawings of Liljevall (Pis. XXXII, fig. no; XXXV, figs. la, b). As to 

 surface angularity produced by median ridges following the radial lines, the first of these 

 figures shows it to a slight extent, while the last does not. The remaining part of the obser- 

 vation was well founded and correct. 



The result of the comparison of Wachsmuth and Springer's specific characters with the 

 facts observed in the two specimens is therefore as follows (the reference being to my 

 figures) ; (1) ornamented surface, present in both; (2) pear-shaped form, in Plate XXXII, 

 figure n but not in the other; (3) surface angularity, in, Plate XXXII, figure n not in the 

 other; (4) straight sutures, in neither. Thus the one character common to both is the 

 ornamented surface, and this applies equally well to /. pyriformis. Upon the second and 

 third characters Plate XXXII, figure n agrees with that species, and for other reasons as 

 well must be referred to it. This leaves only Plate XXXV, figures la, b, which as to the 

 second and fourth characters stated by the authors does not agree with the diagnosis of 

 /. gotlandicus. Nevertheless, as the general form is not wholly constant and cannot be 

 regarded as a predominant character, it seems to me that we are at liberty to select such of 

 the Swedish forms as are ornamented and without surface angularity, and then limit the 

 species by such other characters as we have found to be important and can define by means 

 of later observations. 



As the ornamented surface is also shared by /. pyriformis we must look for characters 

 differentiating that species, and when by this means /. pyriformis is detached, that which is 

 left may be taken as /. gotlandicus. Such a character may be found in the development of 

 the basals, bringing Plate XXXII, figure n under /. pyriformis on account of their large 

 size, aided by the presence of four IIBr as against three. There will thus remain the 

 Swedish specimens which have the basals very small, visible if at all in a side view only as 

 mere points, and usually three IIBr; with these characters Plate XXXV, figure la, b alone 

 out of Angelin's specimens agrees, and it may therefore be taken as the type of a very well- 

 defined species, of which we have now a number of other characteristic specimens from 

 Gotland. Thus it results that of Angelin's figures under the name Ichthyocrinus laevis, his 

 plate 9, figures iya-c = I. pyriformis (PI. XXXII, figs, ua-g herein) ; and his plate 22, 

 figures 20, 21=/. gotlandicus (PL XXXV, figs, la, b, herein). 



Mr. Liljevall has drawn for me all the specimens that are known at Stockholm, includ- 

 ing two views of Angelin's plate 22, figures 20, 21 (PI. XXXV, figs. \a, b herein), and the 

 beautiful enlarged figure of a fragment (fig. 16) showing the manner in which the distal 

 ends of the arms are coiled, proving that the arms were much longer, and afforded more food- 

 gathering surface, than appears from the usual condition of the specimens. 



The crown is a low, broadly swelling ovoid, with side's strongly convex, except for a 

 very short reverse curve in some specimens just above the base. The average of seven 

 Gotland specimens gives a proportion of height to width of 1 to 1.6 at the upper IIBr, which 

 is about that of /. corbis; but it is in marked contrast to the form of /. pyriformis, where the 

 same dimensions are as 1 to 1.2. All these specimens have also quite uniformly three secundi- 

 brachs, as contrasted with the usual four in the Swedish and English specimens of 

 /. pyriformis. 



It is now most interesting to observe in the six Swedish specimens of which we have 

 the basal view (PI. XXXV, figs. 2-6, 15), how the peculiar construction of the base cor- 

 relates with these two characters and differs from that of nearly all other species. The basals 

 are very small, occupying little over half the space covered by the top stem ossicle ; only 

 occasionally does the narrow tongue of one of them reach to the outer surface so as to be 



