266 



NATURE 



\_7an 20, 1 88 1 



that even with joints all made by fusion of the glass it was well 

 nigh impossible to get rid entirely of hydrogen. Mr. Crookes 

 has, I believe, found that the last traces of moisture adhering to 

 glass can only be expelled by heating to the softening point of 

 the glass. This tallies with my own experience. In a series of 

 experiments on the ultra-violet water spectrum I had occasion to 

 photograph the spectra of sparks in sundry gases wet and dry, 

 and found that in gases which had been passed through a tube 

 full of phosphoric anhydride the water-spectrum still appeared 

 strongly. Even when the gas had been passed very slowly 

 through two tubes each half a meter long filled with calcium 

 chloride, and then through a similar tube full of phosphoric 

 anhydride, a'ld the part of the tube where the wires were sealed 

 had been heated strongly for a long time, while the current of 

 gas was passing, traces of the water spectrum still often ap- 

 peared. But Dr. Watts did not see the hydrogen lines in his 

 tube. My difficulty has always been to avoid seeing them A\'hen 

 the pressure of the gas was sufliciently reduced and a large 

 condenser used with the induction coil. True : tubes of gas may 

 not always show them even when hydrogen is known to be 

 present. The spark takes a selected course of its own, and does 

 not always light up all that is in the tube. Carbonic oxide does 

 not generally show oxygen lines, and in tubes exhausted by a 

 Sprengel pump the lines of mercury do not usually appear until 

 the pumping has been carried far. A real test would be to see 

 whether when the spark gives the line-spectrum of carbon the 

 hydrogen lines do not also appear. The experiment with naph- 

 thaline Prof. Dewar and I have repeated and di-cusscd else- 

 where, so I will say no more on it than this, that parity in 

 regard to chemicals is a relative rather than an absolute quality, 

 and that it is only from a long series of experiments chosen with 

 a view to eliminate the effects of .accidents of all kinds that any 

 safe induction in this kind of spectroscopy can be reached. 

 Cambridge, January 4 G. D. Liveing 



[To save time we submitted Prof. Liveing's letter to Mr. 

 Watts, who sends the following reply. — Ed.] 



I SEE no reason why india-rubber stoppers may not be used in 

 the construction of an apparatus to be filled with a gas at atmo- 

 spheric pressarc, or nearly so. The ca^e would be altogether 

 different if we were concerned with the construction of a vacuum 

 tube, and I take it that most of these statements of the difficulty 

 of getting rid of the last traces of moisture and of hydrocarbons 

 adhering to the glass refer to cases where the pressure is to be 

 only a few millimetres. But when a current of cyanogen at 

 atmospheric pressure, made from dried mercuric cyanide, is 

 passed through a U-tube filled with phosphoric anhydride, the 

 gas is surely dry to all intents and purposes (I do not say that the 

 glass would not give off traces of moisture, &c., if the pressure 

 were to be reduced to an extreme point) ; at least there can be 

 so little hydrogen present in the tube that to ascribe the spectrum 

 given by the tube to the hydrogen present in it is to adopt an 

 extreme hypothesis, ^^•hlch must be supported by cogent experi- 

 mental evidence before it can be accepted. 



But if the defect of the experiment be in the use of india- 

 rubber there can be no great difficulty in constructing the appa- 

 ratus entirely of glass, and if we are to give up the view that the 

 groups 8 (5165 to 50S2) and 7 (5635 to 547S) are due to carbon, 

 it must be shown that they are not present in the spectrum of 

 the spark in cyanogen at atmospheric pressure' when sufficient 

 precautions are taken to obtain the gas pure. I have never 

 examined the spectrum of the spark in cyanogen withriut seeing 

 them, and have every confidence that Prof. Liveing will still 

 find them there after lie has taken all the precautions he may 

 think necessary. 



But admitting for the sake of argument the justice of Prof. 

 Liveing's contention that the cyanogen in ray experiment con- 

 tained a trace of hydrogen and that the naphthalin contained a 

 trace of nitrogen, then this seems to be the theory offered for our 

 acceptance — that the spark in nitrocarbon gas containing a trace 

 of hydrogen gives the lines of hydrocarbon, and that the spark 

 iu hydrocarbon gas containing a trace of nitrogen gives the lines 

 of nitrocarbon. Does Prof. Liveing h.(A&.both of these hypotheses 

 to be reaionible? W. M. Watts 



Geological Climates 



inc. icucr or Prof. Haughton in last wecK s ivature so 

 bristles wiih figures and calculations that some of >our readers 



may feel a little puzzled and may be unable to detect the fallacies 

 that lurk among them. The question is far too large a one to be 

 fully discussed in your columns. I shall therefore confine myself 

 to pointing out the erroneous assumptions and f.alse inferences 

 which vitiate all the learned Professor's calculations, having done 

 which my own theory will remain, so far, intact. 



The whole argument against me is based upon an ** ideal ice- 

 cap," extending from the Pole to lat. 60°. A considerable but 

 unknown thickness is given to this imaginary field of ice, and it 

 is then calculated that the three great ocean streams, even if 

 admitted to the Arctic area in the manner I suggest, would not 

 get rid of this mass of ice. There are however several important 

 misconceptions and illogical deductions underlying the whole 

 argument, and when these are exposed the results, however 

 accurately worked out, become completely valueless. 



We first have it stated that if heat and cold were uniformly 

 distributed over the Polar regions the whole would be per- 

 manently frozen over, and an ice-cap be formed of great but 

 varying thiclaiess, diminishing from the Pole to about lat. 60°. 

 But even this preliminary statement is open to serious doubt ; for 

 ice cannot be formed without an adequate supply of water, and 

 over a large part of the Polar area no more snow falls than is 

 annually melted by the sun and by warm southerly winds blowing 

 over the heated land-surfaces of Asia and America. Admitting 

 however that any such ice-cap could be formed, it would certainly 

 not form in one year but by the accumulations of a long series of 

 years ; and any estimate of the total heat required to melt it has 

 m bearing whatever on the annual amount that would be 

 sufficient, since this depends solely on the average thickness of 

 the ice annually formed, of which Prof. Haughton says nothing 

 whatever. 



The amount of rainfall in the Arctic regions (mostly in the 

 form of snow) is certainly very small. It is estimated by Dr. 

 Rink to be only twelve inches in Greenland, and this is probably 

 far above the average. All that falls on the inland plains of 

 Asia, Europe, and America is however melted or evaporated by 

 the action of the sun and air far from the influence of the Gulf 

 Stream. The thickness of ice formed annually over the whole 

 area of the Arctic Ocean I have no means of estimating. In 

 open water in very high latitudes it may be considerable, but 

 perennial ice-fields can only increase very slowly. I should 

 therefore very much doubt if the thickne-s of ice now formed 

 annually over the whole Arctic area averages nearly so much as 

 five feet ; and Prof. Haughton himself calculates that our own 

 Gulf Stream is now capable of melting this quantity. 



The first assumption, therefore — that the amount of heat 

 required to be introduced into the Arctic regions in order to raise 

 their mean temperature above the freezing-point is "accurately 

 measured " by the amount required to melt an "ice-cap " covering 

 the whole area to a thickness of several hundred feet — is grossly 

 erroneous ; and it is so because it takes the hypothetical accumu- 

 lated effects of many years Arctic cold under altogether impossible 

 conditions, and then estimates the amount of heat required to 

 melt this whole accumulation in one year ! 



But we find a second and equally important error, in the 

 assumption (involved in all Prof. Hauuhton's arguments and 

 figures) that all the 'ice of the alleged " ideal ice-cap " must be 

 melted by that portion of the Gulf Stream which actually enters 

 the Polar area, where its temperature is taken to be 35° F. or 

 only 3° above the melting point of ice. A large quantity of the 

 Arctic ice, however, even now floats southward to beyond lat. 50° 

 in both the Atlantic and Pacific, and is melted by the warmer 

 water and atmosphere and the hotter sun of these lower latitudes. 

 Now, as it is an essential part of my theory that much of Northern 

 Asia and North America were under water at those early periods 

 when warm climates prevailed in the Arctic regions, it is clear 

 that whatever Arctic ice was then formed would have a freer 

 passage southwards, and as the south-flowing return currents 

 would then have been more powerful and more extensive than at 

 present, a much larger proportion of the ice would have been 

 melted by the heat of temperate instead of by that of Arctic seas. 



Prof. Haughton admits that the Kuro Siwo and the Mozambique 

 currents together, if they entered the Polar seas, would be equal to 

 the melting of a layer of ice inore than thirteen feet thick over 

 the whole area down to lat. 70°. But if our own Gulf Stream 

 is sufficient to get rid of the wdiole of the ice that now forms 

 annually — as Prof. Hau^jhton's figures show that it would pro- 

 bably be, and as it would be still more certainly were Greenland 

 depressed, th.is ceasing to be the great Arciic refrigerator and 

 ice-accumulator — then the heat of the other two currents would 

 be employed in raising the temperature of the Arctic seas above 



