576 



SCIENCE. 



[N. S. Vol. V. No. 119. 



10 11 12 13 14 K 16 17 18 19 JO 21 23 23 24 25 

 Thousands of B.T.U. and pounds of Steam per H.P. per Hour 

 — Efficiency Curves, ideal and Actual, 



where the value of a varies from 18 on the 

 line A, to 25 on B, 30 on C, and to 22 and 

 24 on lines F and E; w being weight of 

 steam per h. p. per hour, p steam pressure. 

 Of the figures representing efficiency, as 

 here recorded, it is prohable that those on 

 the line C may be accepted as accurate. 

 Those obtained with reheaters in use are 

 obviously less certain, and may be subject 

 to some error. On the whole, the writer 

 considers that the assignment of the lines 

 E and F as those to be attributed to suc- 

 cessful practice with large engines, and as 

 representing the ' promise and potency ' 

 of high-pressure steam, is well justified. 



R. H. Thueston. 

 CoENELL University. 



TEE ORIGIN OF THE TEETH OF THE MAM- 

 MALIA. 



Professor H. G. Seeley, F.R.S., in a 

 series of memoirs in the Philosophical Trans- 

 actions, during the past three or four years, 

 has been describing the Upper Triassic 

 vertebrates of South Africa. Certain of these 

 animals are upon the border line between 

 the Reptiles and Mammals, and, as Profes- 

 sor Seeley points out, show a most remarka- 



ble intermingling of characters. The cranial 

 characters, with the exception of the paired 

 occipital condyle, are mainly reptilian; 

 the dental characters, and this is the point 

 to which I wish to especially draw atten- 

 tion, are pro-mammalian. The point of 

 particular interest is that within this group 

 are found all the primitive mammalian types of 

 teeth. Lycosaurus is haplodont ; Galesaurus 

 Cynognathus, both members of the Cynodon- 

 tia or Carnivorous division, are tricono- 

 dont. The teeth are as clearly divided into 

 incisors, canines, premolars and molars 

 as those of the lower Jurassic mammals. 

 The dental formula approximates that of 

 the stem mammal. These animals parallel 

 or are actually related to the great ' proto- 

 dont-triconodont-trituberculate ' phylum of 

 mammalia, which includes the Marsupials 

 and Placentals. In a distinct division of 

 herbivorous reptiles, which Seeley terms the 

 Gomphodontia, we find a corresponding par- 

 allel or ancestral relation to the ' multitu- 

 berculate ' phylum of mammals, including 

 the Multituberculata and possibly the Mono- 

 tremata. Here, in fact, actually belongs 

 Tritylodon, which upon good grounds has, 

 until recently, been considered a multitu- 



