626 



NATURE 



[Oct. 30, 1879 



them, and that such an extension "cannot therefore have ex- 

 isted." Moreover, the beautiful map of ocean depths with 

 which the volume is illustrated shows a somewhat less depth 

 than 2,000 fathoms on a slightly curved line between the islands, 

 and I believe about the same depth exists between Madagascar 

 and Africa, which have certainly at one time been joined. 



There are some other matters touched upon on which I still 

 venture to differ from my reviewer, especially as to the marvel- 

 lous character of the Easter Island and other remains, and as to 

 the value of the substitution of more for less liberal sectarian 

 teaching in the Sandwich Islands ; but on these points I have 

 quoted authorities of considerable weight, and I leave my 

 readers to form their own opinion. As to all "matters of fact, I 

 gladly accept correction from one who evidently writes with the 

 advantage of a personal acquaintance with most of the countries 

 referred to in his article. Alfred R. Wallace 



Climatal Effects of Eccentricity 



I AM grateful to Dr. Croll for noticing my letter. But I 

 continue to think that if what seems to me to be the fundamental 

 proposition of his theory, and which I quoted at the begin- 

 ning of my former letter, be correct, and if the manner in which 

 he and his reviewer have applied it be likewise correct, 'then 

 we ought to find those differences in the air temperatures which 

 my equations indicate. I say air-temperatures, because in Dr. 

 CroU's theory changes of climate are referred to the varying 

 distance of the sun, and climate depends on the temperature of 

 the air. 



The heating effect of the sun, other things being equal, has 

 been hitherto assumed to be proportional to the excess of the 

 temperature of the place above the temperature of space. A 

 remarkable result which_Pouillet had arrived at, and of which I 

 was not aware when I wrote, shows that this method is incorrect. 

 And I believe that what follows will to some extent afford a 

 reply to the question which I have propounded, and at the same 

 time have a proportionate bearing on Dr. CroU's theory. I 

 quote Pouillet's words from the translation in Taylor's "Scientific 

 Memoirs," vol. iv. p. 83. 



"The total quantity of heat which space transmits in the 

 course of a year to the earth and to the atmosphere .... would 

 be capable of n.elting upon our globe a stratum of ice of 

 26 metres thickness. We have seen that the quantity of solar 

 heat is expressed by a stratum of ice of 31 metres. Thus, 

 together, the earth receives a quantity of heat represented by a 

 stratum of ice of 57 metres; and the heat of space concurs in 

 this for a quantity which is five-sixths of the solar heat. Between 

 the tropics the heat of space is only two-thirds of the solar heat ; 

 for the latter is there represented by a stratum of ice of 39 

 metres." 



These surprising results arise from the unequal absorption 

 exercised by the atmosphere upon the heat rays proceeding from 

 the stars and from the earth respectively. 



It appears then, that, in applying Dr. CroU's proposition, we 

 ought not to use the value of the temperature of space in forming 

 our proportion, but we ought to use the temperature which the 

 surface of the ground would assume were the sun extinguished. 

 This Pouillet puts at - 89°, or - 128° F. The substitution of 

 128 for S., instead of 239, reduces my calculated difference 

 between the January and July temperatures at the equator to 

 n" F., i.e., by about one-half. 



If we make the same correction in the case, the high eccentri- 

 city at aphelion, for which the Quarterly Revir.ver has calculated 

 the January temperature of England, and found it 3° F. (I make it 

 even lower), the new temperature comes out 17° F., which can 

 hardly be thought low enough to cause any extreme difference 

 from the present climate. O. Fisher 



October 25 



The statement quoted by Mr. Fisher from Dr. Croll (Nature, 

 vol. XX. p. 577) that "the temperature of a place, other things 

 being equal, is proportional to the heat received from the sun," 

 is based on the assumption of Newton's law of cooling, viz., 

 that the rate of cooling of a body is proportional to the excess of 

 its temperature above that of the surrounding medium. This is 

 approximately true only when the excess is small. When the 

 excess becomes large the rate of cooling augments much more 

 than in proportion. The amount of heat which must be supplied 

 to a body in order to maintain it above the temperature of the 

 surrounding medium is proportional to what would be its rate of 



cooling. Hence this amount increases as the excess of tempera- 

 ture increases — proportionally while the excess remains small, 

 but much more than proportionally when it becomes large. 

 Conversely, the temperature increases more slowly than the 

 amount of heat supplied, and any variation in the supply will 

 affect the temperature produced in a degree which is less for a 

 large excess than for a small one, and, therefore, less than 

 Newton's rule would give. The excess of the earth's mean 

 temperature above that of space is large, and consequently 

 calculations of changes based on Newton's rule must be in excess 

 of the truth. 



The formula obtained by MM. Dulong and Petit (Stewart on 

 " Heat," Art. 235) from the rate of cooling of a thermometer, 

 bulb in vacuo makes the necessary supply of heat proportional to 

 ( I '0077' — I ), where t is the excess of temperature in Centigrade 

 degrees. If we apply this to the case Jof the earth, and take 

 80" F. as temperature at the equator when the earth is at its 

 mean distance from the sun, then the resulting temperatures at 

 its greatest and least distances with our present eccentricity, are 

 given as about 74° and 85° respectively. The fluctuation, which 

 Mr. Fisher makes 21°, is reduced to about 11°. The fall in 

 temperature which would follow a stoppage of the Gulf Stream 

 is made by Newton's rule 59° (" Climate and Time," p. 36) : the 

 more accurate formula reduces this to about 37°. Dr. Croll 

 suggests that the temperature of space may be lower than ii 

 usuaUy assumed (p. 37). If it be taken as iiiisolute zero ( - 459^ 

 F.) the fall would not even then come out much greater than 

 45° F. 



Several of Dr. CroU's tables should be sumlarly modified ; at 

 the same time it would be scarcely correct to say that these 

 changes "touch Dr. CroU's theory somewhat closely." They 

 do not invalidate the general contention, that a diminution of 

 the Gulf Stream must diminish the meantemperature of northern 

 regions to a very serious degree. E. HiLt 



St. John's College, Cambridge, October 25 



The Weather and the Sun 



Prof. Piazzi Smyth in his communication to Nature, 

 vol. XX. p. 431, evidently infers that changes in the condition of 

 the sun must needs affect every part of the earth in the same 

 way, whereas we have many meteorological analogies, which 

 favour the notion that totally opposite effects may arise in different 

 parts of the earth from the action of the same primary causes. 

 For example, it is generally assumed that the same tropical heat 

 which gives the primary impulse to the desiccating north-eist 

 trade wind of sub-tropical latitudes, furnishes the energy which 

 exhibits itself in tlie almost constant precipitation under the 

 equator. Any variation in the degree of this heat, shoidd 

 consequently affect localities situated in the region of the trades, 

 and the equatorial calm-belt, in a diametrically opposite manner. 

 Moreover, the notion that the British and Indian rain falls vary 

 together now is altogether inconsistent with the well-known want 

 of similarity between them, both as regards seasonal distribution 

 and annual quantity in the past. It is also remarkable that 

 while the present deluge both here and in India is traced to the 

 sun's '^recovering his forces and beginning already to shine after 

 his recent languid spotless years with increased radiation on the 

 great oceans of the south," the rainfall of England between lati- 

 tudes 50° and 55" N. reached a decided maximum in 1877, the 

 year when the sun was, to adopt the favourite metaphor, affected 

 with the most extreme languor, and has been very high all 

 through the period of unusually marked spot minimum, from 

 which we are but just emerging. 



The following figures from Mr, Glaisher's reports will iUustrate 

 what I have just said. 



Great Britain, Lat. 50° 



Years. 

 187s 

 1876 

 1877 

 1878 



-55° N. 



Rainfall in inches. ' 



34"04 

 34-60 



... 38-55 

 3261 



More valuable results will generally accrue to science if, instead 

 of founding vague hypotheses on a fancied likeness between iso- 

 lated weather conditions, at places where the prime meteorological 

 factors act in a totally dissimilar manner ; induction is made from 

 results which are derived from trustworthy data, and anticipated 

 by a knowledge of admitted physical principles. As an example 

 of this latter kind, allow me to conclude this letter by exhibiting 



