45S N. IF. Willixtoit — Restoration of JJinnoseelis. 



forty-one in specimen 809, and twenty-six in specimen 811. 

 Specimen ~No. 650 of the Chicago University collections com- 

 prises about twenty caudal vertebrae, for the most part con- 

 nected in a basal series. Unfortunately in none of these 

 specimens are the spines quite complete throughout. From 

 a comparison of the four different specimens, however, most 

 of these have been determined with certainty, as shown in the 

 drawing. In specimen 809 three of the proximal chevrons are 

 preserved complete in natural articulation ; in specimen 811 a 

 number of the distal ones are also connected with the centra; 

 while those present in 819 are shown in fig. 25. The three 

 specimens of the Yale collection are almost identical in size, 

 819 being perhaps a trifle larger than the others, though not 

 much, the total length of the first twenty-three caudal verte- 

 brae being about three-fourths of an inch greater than that of 

 the same vertebras of 809. The Chicago specimen, however, 

 is distinctly smaller, and also presents some slight differences 

 which might perhaps be accounted of specific value did we 

 know what specific characters are in these old reptiles. 



I give herewith a number of figures of some of the best pre- 

 served parts of these different specimens, all one-half natural 

 size. The humeri (figs. 1, 2, 31) are best preserved in the 

 Chicago specimen; in the Yale specimen J^o. 811, from which 

 my previous figure and descriptions were made, the ectocon- 

 dyle had evidently been crushed somewhat inward from its 

 normal position ; it really projects nearly at right angles 

 dorsad from the distal plane of the bone, much as in Diadectes 

 and Diasparactus. The ectepicondyle is larger and better 

 preserved than in- any of the Yale specimens of this bone, and 

 all the processes are better defined. Perhaps the clay matrix 

 has had something to do with the better preservation of this 

 specimen. The radius of the Chicago specimen (figs. 4, 5) in 

 comparison with that of 811, Y. U. (fig. 28), seems to be a little 

 less stout. A well preserved femur and the fibulas of the 

 Chicago specimen are shown in figs. 6-9, the lower end of the 

 femur completed from that of the opposite side. Especially 

 characteristic of Limnoscelis, in which it agrees with Sey- 

 mouria, but disagrees with the Diadectidw, are the large size 

 and low position of the trochanter, and the high, thin adductor 

 crest, as in most of the contemporary amphibians. Both the 

 tibia and the fibula of this specimen are a little less stout, and 

 less expanded at their ends than in the holotype. Two views 

 of a typical posterior dorsal vertebra of this specimen are 

 shown in figs. 15, 16. In my original description of the genus 

 I gave as a characteristic the presence of an infracentral fossa 

 on the presacral vertebras. I am now not sure whether this 

 character is of importance, since it seems to be absent in the 



