( 81 ) 



Cahalus dieffenbachii there .are only four such vertebrae, but Ocydromus resembles 

 the fossil in this respect. Behind the interacetabular fossae come three vertebrae 

 with long slender processes ; in the smaller rails there are usually only two. Behind 

 these again there are two vertebrae with short stout transverse processes abutting 

 against the inflected border of the ilium. There are in all fourteen vertebrae in the 

 " sacrum." As in other rails, the posterior renal fossae form large pocket-like recesses, 

 partly floored by an inflected portion of the ilia. 



The dimensions (in millimetres) of six specimens of the pelvis were : — 



In the femur the curvature characteristic of the rails is well markeS. The only 

 points in which it differs from that of Ocydromna are — (1) it is somewhat stouter 

 in proportion to its length ; ('!) at the lower end of the linea aspera there is a short 

 cristiform ridge, running obliquely to the median angle of the inner condyle, and 

 forming the inner border of the comparatively shallow popliteal fossa (this crest 

 is also wanting in Aphawipteryx and Erythromachus, but is well developed in 

 Fidica) ; (3) the surface for the fibula does not form a groove, but is a mere shelf. 

 The points of attachment of the sling for the tendon of the biceps cruris are well 

 marked, as are all the muscle impi-essions. 



The femora of Aphaiiapteryx and Erythroriiaclias also closely resemble the fossil, 

 but judging from the figure, the former appears rather straighter. In Aptornis 

 the femur is proportionately considerably stouter, is nearly straight, and has a dee)) 

 popliteal fossa. 



The length of this lione ranges from about 90 to 106 mm., the largest specimen 

 thus being roughly 18 per cent, larger than the smallest. The average size of 

 thirty-three specimens was 97'-l mm. ; in a specimen of that size the width of 

 the distal end is 23 mm., of the proximal end 22 mm., and of the middle of the 

 shaft S"5 mm. 



The tihio-titrsus is extremely like that of Aphaiiapteryx broecki, and con- 

 sequently, as ]\Iilne-Edwards has shown in his valuable paper on that bird, it is very 

 similar to the tibia of Ocydromtis. The shaft, however, appears to be somewhat 

 straighter, and the extremities, particularly the lower, more massive; the upper 

 articular sm-face is proportionately wider from side to side, and the outer and inner 

 glenoid facets are separated behind by a wider notch. As in Aphanapteryx, there 

 is a deep fossa between the inner glenoid facet and the base of the anterior cnemial 

 crest; this latter rises above the articular surface to a somewhat greater degree than 

 in the Mauritian bird. 



The channel for the tendon of l\w jjeroueus profundus is only slightly marked, 

 the retinaculum remaining unossified in all the specimens I have seen. The fibular 

 crest is prominent, and the lower end of the fibula reached down to about the 

 junction of the middle and lower thirds of the bone. The length of this bone varies 

 from 122 to 140 mm., the longest specimen measured being about 1 J per cent, longer 

 than the shortest. The average length of twenty-eight specimens was 130 mm. In 



6 



