Nov. 19, 1885] 



NATURE 



53 



The paragraphs are as follows : — 



Mr. Baldwin Latham, in a discussion wliich ensued upon the 

 Report of the Committee on Decrease of Water Supply 

 {Quarterly Jotirnal Roy. Met. Soc, p. 22:j), said : — 



" The records showed that there appeared to be a recurrence of 

 low water every ten years. There was lower water in 1824 and 

 in 1835 ; the period 1844-5 was low, especially when compared 

 with the years immediately before and following ; 1854 was re- 

 markably low ; also 1S64-5, 1874-5, ^'^^ "ow they come to the 

 present low period of 1884-5. 



" As to what was the cause of this marked periodicity it was 

 very desirable to ascertain, and, having pointed it out, probably 

 some light might be thrown on the subject." 



The other is from the American Meteorological y oiirnal 'anie.x 

 the heading "Cold Winters in Michigan," and the writer 

 says : — 



" It is interesting in this connection to notice that the local 

 reports of extremely cold winters place them at intervals of be- 

 tween ten and eleven years. . . . The winter of 1842-3 is thus 

 shown to have been extremely cold ; also the winter of 1853-4 ; 

 the winter of 1863-4 noted for its terribly cold new year ; 

 the winter of 1874-5, when the frost penetrated into the ground 

 in Port Huron four to six feet, there being scarcely a thaw 

 between January i and the middle of March ; and, lastly, the 

 winter of 1884-5, which beats the record for extreme cold 

 during January and February." 



I may add that before I had seen either of these paragraphs 

 I had concluded from other sources that the years 182 1 -2-3-4, 

 1833-4, 1844-5, 1866-7, and 1875 -7 were characterised by mild 

 winters in Europe and unusual cold in Iceland and America, 

 being preceded in most cases by drought during the summers ; 

 but of course this represents merely the result of a preliminary 

 glance at some reneral records of noteworthy seasons. 



November 9 E. Douglas Archibald 



photography of the Corona 

 I HAVE been following with interest the communications 

 which have been made from time to tim; to Science by Mr. 

 W. H. Pickering regarding the photography of the corona in 

 full sunshine. Whilst admiring the manner in which he has 

 built up his theoretical objections to its possibility, I am 

 forced to dissent from his deduction from the fact that the 

 theory does not fit in with the results actually obtained during 

 the eclipses observed in Egypt and the Caroline Islands. I 

 iiave in my hands at present spectrum and other photographs 

 of the corona made during the expeditions to those localities, 

 and from them I gather he has evidently much underestimated 

 the photographic brightness of the corona as compared with 

 that of the sky. As I propose shortly to read a paper before the 

 Royal Society on the subject, I cannot enter into details at the 

 present moment. K\\ I will say is that the comparative photo- 

 graphic intensity of both can be estimated with approximate 

 exactness from the data I have by me. 



I write this for insertion in your columns, as in your last issue 

 you have a note regarding Mr. W. II. Pickering's communica- 

 tions on this subject. " W. de W. Abney 



Permanence of Continents and Oceans 

 Many naturalists are accustomed, in lecturing, to speak of 

 the existing ocean basins as "permanent." Though this must 

 to a large extent be a true statement, many geologists at all 

 events must be perfectly aware that the former distribution of 

 life requires that nearly all lands however remote at present, 

 must have been, perhaps more than once, in connection with 

 each other. Tropical South America is perhaps the most iso- 

 lated continental province now existing. I would ask these 

 naturalists to explain how its species of tropical genera not 

 peculiar to it got there, and how many of them came to be 

 represented in Europe in Tertiaiy times. 



That the lands are always chiefly centred about the same 

 spots, and also the converse, would, I think, be an acceptable 

 way of putting it ; but that the Atlantic was never bridged 

 except towards the Arctic and Antarctic circles, is a statement 

 that is unwarrantable because contradicted by unimpeachable 

 evidence. J. Starkie Gardner 



History of Elasticity 



In order to estimate Poncelet's services to the theory of elas- 

 ticity I am extremely desirous of examining certain works by 



him. These works are not to be found in the London or Cam- 

 bridge Libraries, and the Paris booksellers to whom I have 

 applied despair of being able to procure copies. It will hardly 

 be possible for me to go to Metz to examine them before the 

 publication of the first volume of the " History of the Mathe- 

 matical Theories of Elasticity." Possibly some of your readers 

 may know of the existence of accessible copies in this country. If 

 so, I should esteem it a great favour if they would communicate 

 with me at University College. 



In 1827-29 Poncelet gave at Metz a " Cours de Mecani^ue 

 Industrielle aux Artistes et Ouvriers Messins." In this "Cours" 

 various important points of theoretical elasticity were considered 

 for the first time. 



It was published as follows : — 



(a) Part I. Lithographed edition, Metz, 1827. 



(/') Part II. First edition lithographed 1828 ; second edition 

 lithographed 1831. 



(c) Part III. Lithographed edition, 1831. 



[d] Part I. First printed edition, Metz, 1829 ; second printed 

 edition, Metz, 1831. 



It is needful to remind your readers that there are numerous 

 other works entitled " Mecanique Industrielle," by Poncelet, 

 published at Liege, Paris, and Brussels, differing from each 

 other, and entirely from the above. These I have examined, 

 but they do not contain what I require. Karl Pearson 



University College, London, November 15 



The Heights of Clouds 



In the very favourable notice of our " Mesures des Hauteurs 

 et des Mouvements des Nuages," in N.ature of October 29 

 (p. 630), there exists a misunderstanding as to the probable 

 errors of our measurements, which makes our observations seem 

 much more inexact than they really are. I therefore ask your 

 permission to correct it. 



Mr. W. de W. A. says : " Perhaps one of the most easily- 

 observed clouds is the cumulus, and we find from a table given 

 that \\\Q. probable error of ohserva'.ioii is very considerable." But, 

 in fact, what is there referred to as an error of observation is not 

 such an error; it is the probable uncertainty ("incertitude 

 probable") depending on tiie variability of tiie pJieuonienon itself.^ 

 This is expressly stated in the treatise. On p. 39 (that following 

 the table quoted) there may be read : *' L'incertitude calculee 

 comprend ainsi et celle dependant de la variation des hauteurs 

 des nuages, et celle provenant des erreurs d'observation. Celle-ci 

 est cependant assez petite par rapport a la premiere et i pen 

 pres constante pour les differentes heures du jour, comme on le 

 trouvera en la calculant separement a I'aide des erreurs moyennes 

 /«." That mean error m is just the mean error of observation in 

 the height of a cloud, and in our " list of observations " we have 

 given it for every observation, as well as the corresponding mean 

 angular error of the observation. By calculating the probable 

 error of an observation of cumulus by means of those values of 

 mean errors we have found it to be 35 metres (instead of 74S 

 metres, as Mr. W. de W. A. thinks it to be), and the probable 

 error of the mean is found to be 3 metres (instead of 40 metres), 

 the whole number of observations being 134.- 



Thus the above assertion is fully justified, viz. that the errors 

 of observation may be quite neglected when compared with the 

 uncertainty depending upon the variability of the heights of the 

 cumuli from one cloud to another. That variability is itself a 

 phenomenon worthy of investigation, varying as it does accord- 

 ing to the hour of the day and the barometrical state of the 

 weather, and that is the reason for which we have calculated it. 

 As to the mean angular error in observing a cloud, we have 

 found it very often to be inferior to that obtained in observing 

 the centre of the sun, and in all the better observations that 

 error is fully comparable to the error in observing the sun, as 

 may be seen from our treatise. This proves that, for such 

 observations, the uncertainty of an identical point of cloud did 

 not exist at all, the whole uncertainty depending on the unavoid- 

 able instrumental errors. Still it may be that the errors are 



■ For the figures in the table quoted represent simply the probable differ- 

 ence of an observation (of the mean found) from the true mean calculated by 

 the method of least squares. 



' For the higher clouds, as the pure cirri, this error w.-is often very great 

 indeed, but it was so because their distance was much too great when com- 

 pared with our basis, the parallax obtained being not greater than 1° or 2°. 

 This year {18S5) the measurements are regularly pursued from the ends of a 

 basis of 1302 metres, and we can now determine with great accuracy the 

 height even of the most elevated cirri, as well as their horizontal and vertical 

 velocities. 



