July 21, 19 10] 



NATURE 



bringing cold water from Baffin's Bay as a surface current 

 round Newfoundland and down the coast to Cape Hatteras 

 and even beyond it. The principle was the same as that 

 which moved Humboldt to attribute the cold water, which 

 we have described in connection with the Pacific coast of 

 tropical South America, to a surface current from the 

 Antarctic Ocean. In the p^per on similarities, &c., above 

 referred to, I showed that Humboldt's explanation 

 postulated an impossibility. The deeper layers of the 

 water on the coast itself are capable ai supplying, as and 

 when required, much more cold than is wanted, and that 

 with the least expenditure of energy. The same is the 

 case with the "cold wall." BesMes the south-westerly 

 winds of the North Atlantic, and perhaps independently of 

 them, the Gulf Stream itself, pouring its waters in a 

 stream of great momentum past the American coast and 

 out into the open ocean, performs the function of a colossal 

 jet-pump, carrying water away from the surface and 

 leaving its place to be taken by the other water which can 

 get there most easily. This is the cold water of the 

 deeper layers in situ. It is this hydraulic cold-water 

 service which tempers the climate of the eastern States. 

 The labours of the U.S. Coast Survey during the last 

 seventy years have shown that fluctuations, both regular 

 and irregular, occur in the flow of the Gulf Stream. 

 These necessarily react on the supply of cold water drawn 

 from the deep and spread over the continental shelf. Such 

 variations are probably the source of the accidents which 

 occasionally occur and cause the extinction of life over 

 large tracts of shoal water on that coast. 



J. Y. BtJClIANAN. 



REMNANTS OF THE FAST. 

 A/r UCH interest attaches to a paper by Mr. R. S. Lull, 

 •'■■'■ published in the " Proceedings of the Seventh Inter- 

 national Zoological Congress, Boston, 1907 " (issued 1910), 

 on the evolution of the horned dinosaurs, or Ceratopsia. 

 .•\lthough early ancestral forms are at present unknown, 

 it is probable that the group took origin from an iguano- 

 dont stock. The earliest known types are Monoclonius 

 and Ceratops of the Judith River beds, the single repre- 

 sentative of the former being the more primitive, and 

 probably ancestral to all the rest. In Monoclonius the 

 orbital horns are much smaller than the nasal one, but 

 in one species of Ceratops the two have become subequal ; 

 both genera show large vacuities in the cervical flange of 

 the skull, which was probably internal. Between the 

 Judith River and Laramie formations occur certain marine 

 formations yielding no dinosaurian remains, but in the 

 basal Laramie occur Agathaumas, of which the skull is un- 

 known. Higher up this is succeeded by Triceratops, in 

 which the vacuities in the cervical flange are obliterated, 

 while in the various species may be traced a gradual in- 

 crease in the size of the orbital at the expense of the nasal 

 horn, the latter becoming almost obsolete in T. elatiis, while 

 it has disappeared in Diceratops, which forms a side-branch 

 by itself. The remarkable genus Torosaurus of the Upper 

 Laramie, although having developed large orbital horns at 

 the expense of the nasal one, retains the long, straight 

 skull, with a large vacuity in the cervical flange, of the 

 Judith River Ceratops monatiis, from which it may be 

 directly descended. Physical changes in their environment 

 seem, in the author's opinion, the most probable cause of 

 the extinction of these marvellous reptiles. 



In the April number of the American Journal of Science 

 Mr. F. Loomis describes the complete skeleton of a new 

 species of the camel-like genus Stenomylus from the 

 Harrison beds of Nebraska. The genus differs from other 

 Tertiary types by the hypsodont character of the dentition. 

 This is considered by Mr. Loomis as an indication that 

 Stenomylus differed from its relatives in habits. The 

 early tylopods of the Protomeryx type probably fed on a 

 mixed diet, while the members of the long-limbed Oxy- 

 dactylus group may have subsisted on leaves and shoots, 

 both retaining the original brachyodont dentition. Steno- 

 mylus, on the other hand, seems to represent a separate 

 branch derived from the ancestral Poebrotherium, which 

 developed a hypsodorft dentition, and took to feeding on 

 hard-stemmed grasses growing on open, arid plains. 



Dr. A. E. Ortmann contributes to the .\pril number of 

 the American Naturalist an article on the theory that a 



NO. 2125, VOL. 84] 



connection between Africa and South America persisted 

 into the Tertiary. According to the Archelenis theory, as 

 originally proposed by Dr. von Ihering, an ancient con- 

 nection between the above-named continents was the last 

 remnant of the much greater equatorial land-mass known 

 as Gondwanaland, an area which was broken up at various 

 dates, and remnants of which are represented by Australia, 

 India, Africa, and Brazil. The separation of Brazil from 

 Africa was the final stage in the dismemberment of the 

 old continent, and it is generally considered that this took 

 place towards the close of the Mesozoic epoch. A study of 

 the Tertiary flora of Patagonia has, however, induced Dr. 

 von Ihering to believe that Archelenis persisted into the 

 Tertiary. It is argued, however, that the facts cited by 

 von Ihering really lead to just the opposite conclusion, 

 while the existence of marine Eocene deposits in many 

 parts of West Africa is likewise an indication that the con- 

 nection between the two continents had ceased. Accord- 

 inglv, the evidence for a Tertiary Archelenis is considered 

 valueless. 



Vol. vii.. No. 2, of the University of Colorado Studies 

 is devoted to an account, by several aiithors, of the results- 

 of a scientific expedition to north-western Colorado. Ire 

 a paper on plant-remains from the Cretaceous of Mesa 

 Verde, Prof. T. D. A. Cockerell describes and figures a 

 fragment of a branch bearing a remarkable resemblance 

 to the Palaeozoic lycopods of the Ulodendron type. U 

 reallv belongs to an araucarian conifer (Geinitzia reicben- 

 bach'i), but its resemblance to lycopods of an earlier period 

 is highly significant in view of the probability of a real 

 relationship between the two groups. 



NON-FERROUS METALS. 



IN many respects the second volume of tTie journal of the 

 Institute of Metals marks a decided advance on the 

 first volume— an advance which serves as a healthy sign 

 of the continued growth of the institute. Perhaps the best 

 sign of this advance is the inclusion, in the second volume, 

 of^a series of abstracts of scientific and technical literature 

 bearing upon the subjects which come within the scope of 

 the institute. These abstracts fill what has hitherto been 

 a decided gap in metallurgical literature ; they are 

 obviously modelled on the very excellent abstracts of the 

 literature of iron and steel which appeared in the Journal 

 of the Iron and Steel Institute while that journal was 

 under the editorship of the late Mr. Bennett Brough. 

 Perhaps the most serious criticism to be offered on these 

 abstracts is that thev are of too indiscriminate a character, 

 mere descriptive papers of small permanent interest bemg 

 accorded equal space with papers of real importance. 



The original papers, which, with the discussions, occupy 

 the greater part of the second volume, have already been 

 referred to in these pages on. the occasion of the meetmg 

 at which thev were read. It is satisfactory to find that 

 the discussion's show signs of free: and vigorous criticism, 

 and that such criticism seems to be accepted by the authors 

 in a kindlv spirit, even though ' at times the cnticisnis 

 are practically destructive. Thus the first paper (Edwards 

 and Andrew on aluminium-copper-tin alloys) is criticised 

 on the ground that the data published do not afford 

 sufficient insight into the facts upon which the authors 

 base their conclusions. The paper of Prof. Turner and 

 Mr. Murrav, on the volume-changes of the copper-zinc 

 alloys, is also challenged as regards the validity of its 

 conclusions on the ground— apparently justified— that the 

 mere measurement of the longitudinal contraction ot a 

 casting can give no true insight into the volume-changes 

 which accompany the passage of the metal from the liquid 

 to the solid state. More than eighty pages of the volume 

 are devoted to the paper of Mr. A. C. M. Snriith on the 

 elastic breakdown of non-ferrous metals, and although the 

 subject presents certain points of interest, it appears to 

 occupy a good deal more than its fair share of space m 

 a journal not specially devoted to such questions as the 

 best means of measuring elastic constants. The paper 

 however, shows clearly the narrow limits within which 

 Hooke's law is applicable to such metals as copper and 

 aluminium; the latter appears to be particularly unsatis- 



1 The Journal of tlie In'tit.ite of MctaU, vol. ii. Pp. 341- Vol. 

 Pp. xi + 360. Edited by G. Shaw-Scott, Secretary. 



