420 



NATURE 



[Feb. 28, I i 



corroded debris, though hkewise finally deepening, widenin^f, 

 and smoothing the basins in the solid rock. 



The Hills and Hill-groups of Britain have all emerged during 

 the gradual denudation of the country, and owe their prominence 

 to tiie greater durability of their materials as compared with those 

 of the surrounding lower grounds. They thus represent various 

 stages in the general lowering of the surface. In many cases 

 they consist of local masses of hard rod;. Such is the structure 

 of the prominent knobs of Pembrokeshire and of Central Scot- 

 land, where masses of eruptive rock, formerly deeply buried 

 under superincumbent formations, have been laid bare by denu- 

 dation. In connectio'i with such eruptive bosses attention should 

 be given to the "dykes" so plentiful in the north of England 

 and Ireland, and over most of Scotland. In numerous instances, 

 the dykes run along the crests of hills and also cross wide and 

 deep valleys. Had the existing topography existed at the time 

 of their protrusion, the molten basalt would have flowed down 

 the hill-slopes and filled up the valleys. As this never occurs, 

 and as there is good evidence that the dykes are not of higher 

 antiquity than the older Terti.ary periods, we may conclude that 

 the present configuration of the country has, on the whole, been 

 developed since older Tertiary time— a deduction in harmony 

 with that already announced from other independent evidence. 



Escarpments are the steep edges of hills in retreat. The 

 British Islinds abound in admirable examples of all ages from 

 early Palaeozoic rocks down to Tertiary deposits, and of every 

 stage, from the almost unbroken line of cliff to scattered groups of 

 islet-like fragments. The retreat of our escarpments can be well 

 studied along the edge of the Jurassic belt from DorseLshire to 

 the headlands of Yorkshire, likewise in the course of the ed'^e of 

 the Chalk across the island. Not less suggestive are some of the 

 escarpments of more ancient rocks, such as those of the older Palzeo- 

 zoic limestones, the Old Red Sandstone of Wales, the Carboni- 

 ferous Limestone and Millstone Grit of Yorkshire, and the Coal 

 Measures of the Irish plain. Our volcanic escarpments are like- 

 wise full of interest— those of the Lower Old Red Sandstone 

 along both sides of the Tay, of the Carboniferous system in 

 Stirlingshire, Ayrshire, Bute, and Roxburghshire, and of the 

 Tertiary series in Antrim and the Inner Hebrides. 



SUN-GLOIVS AND VOLCANIC ERUPTIONS IN 



ICELAND 

 TN reply to the inquiry despatched to me by Nature with last 

 mail, whether any remarkable sun-glow had been observed 

 recently in Iceland, and which, I learn, has been observed in 

 nearly all parts of tbe world, and whether any volcanic eruption 

 had Lately taken phice in the island to which the same might be 

 attributable, I beg to relate, as regards the first of these points, 

 that on November 23, between 5 and 6 p.m., I noticed for the 

 first time an unusual and striking purple intensity of the sky, a 

 phenomenon which was also oV)served on the subsequent morn- 

 ings and nights. I did not attach much importance to this 

 phenomenon at the time, tnrough the circumstance that I was 

 told that sunrises and sunsets were generally attended by very 

 intense auroras here, and since then I have bad so few opportu- 

 nities of seeing the sky free frooi clouds th.at I have not 

 observed any similar phenomenon. I learn, however, on inquiry 

 here, that the same glow was observed once or twice before 

 Christmas by several persons. On one occasion, January 30, the 

 sky was perfectly clear several hours after sunset, but there was 

 no unusual glow. 



With regard to the second point, as to recent volcanic erup- 

 tions in the island, I have not much new information to transmit 

 (Nature, vol. xxix. p. 343). The only thing we know as to 

 this is that a man has written a letter to an Iceland paper stating 

 that on October 8 and 9 last year he was at a farm about three 

 geographical miles east-north-east inland from the w ell-known 

 fishing village Seydisfjord, on the east coast, when he saw, on 

 the first-mentioned day, in the direction of the unexplored 

 gigantic volcanic mountain, the Vatnajokull, about 130 geo- 

 graphical miles in extent, in the south-eastern corner of the 

 island, two columns of fire, and on the following morning, in the 

 same direction, two columns of smoke. He adds that a similar 

 phenomenon was observed on the farm two days previously. 

 It is also reported to us here that ashes have fallen in 

 Seydisfjord. 



It is most probable that these eruptions have occurred in the 

 same place where similar phenomena have been observed several 

 times in recent years, viz. in the neighbourhood of the Kveric 



Mountains on the mrth side of the Vatnajokull, and that there 

 are, in all probability, several volcanoes in activity in this district, 

 which is utterly unapproachable to explorers. 



There is, however, no reason to assume that eruptions of any 

 magnitude have recently taken place in .any other part of the 

 island, as such an occurrence would soon have been reported by 

 some means or another to us here. 



If, therefore, the remarkable sun-glows of which I read 

 are attributed to terrific volcanic eruptions, the latter must l>e 

 sought in other localities than Iceland. 



SOPHUS Tromholt 



Reykjavik, Iceland, February i (by mail February 8) 



COMPOSITE PORTRAITURE ADAPTED TO 

 THE REDUCTION OF METEOROLOGICAL 

 AND OTHER SIMILAR OBSERVATIONS ■ 



T T has often been remarked that one of the main, if not the 

 chief, of the difficulties the meteorologist has to contend 

 with, is the enormous amount of preliminary labour which has 

 to be expended in the not very pleasing task of forming the 

 observations he may wish to discuss into tables, casting the 

 columns of figures so obtained, and then computing the means. 

 Should, as in many cases nowad.ays, his original material be in 

 the shape of curves, e.g. barograms, thermograms, or anemo- 

 giams, he has first to reduce these to figures by tabulation, 

 before he can attempt any step towards their reduction. 



The deterrent nature of these preliminary operations not 

 unfrequently forms a complete bar to the entering upon most 

 interesting investigations with a view to the advancement of the 

 science, in the case of persons unable to devote sufficient time t ■ 

 such labour, which may almost be termed drudgery. To cite 

 example-, a glance at the recently published papers in the 

 Proceedings of the Royal Society, by Prof. Balfour Stewart (vol. 

 XXV. p. 577) and by Mr. C. Chambers (vol. xxxiv. p. 231), in 

 which they endeavour to trace a possible intimate connection 

 between solar and terrestrial phenomena, will show the immense 

 amount of calculation they had to perform in order to arrive at 

 their results — how, for instance, preliminary means had to l>c 

 taken of three days' observations and the result assumed to I'C 

 a corrected value for the middle day of the three, then, after ihe 

 whole series had been si treated, a .second or even a third set iil 

 averages computed. The author has also a 1 vely recollection of 

 the excessively tedious calculations required to eliminate in a 

 somewhat similar manner the effect of disturbances in the discus- 

 sion of the Kew magnetic observations for the late Sir E. Sabine. 

 With the view of arriving at results by a shorter cut, the auth ir 

 has been led to consider the possibility of employing a method 

 suggested by an examination of the highly ingenious system of 

 composite portraiture invented by Mr. Francis Galton, F.R.S.. 

 and utilised in his anthropological studies. 



Mr. Galton's method of experiment is based upon the fact that 

 certain groups of people possess certain physiognomical features 

 in common. This agree nent of feature is usually characterised 

 by the term " family likeness." In order, the-efore, to select this 

 particular element from the others, and to obtain a picture in 

 which it is most strongly defined : or, in other words, to firm .t 

 characteristic portrait of the grjup of individual-, Mr. Galton 

 employs a series of photographs. These, representing a lar^e 

 number of men or women, are first reduced to the same scale, 

 and then projected successively upon a sensitised photographic 

 plate, having been previously so arranged that the eyes or other 

 s.alient feature shall always fall on the same porti^n of the 

 plate. 



In this manner a negative is eventually obtained which gives a 

 print depicting a countenance which, although resembling 1 ut 

 partially any one of the component portraits, gives a fair typical 

 picture of the group of individuals. Among other results Mr. 

 Galton has detected the likeness existing in various clisses ol 

 criminals, and also in patients suffering from the .same disease, 

 as well as the more marked features transmitted through the 

 different members of a family. 



Since in meteorological investigations the desire is to select 

 and to identify the one particular variable running through a 

 group of phenomena, it has appeared to the author, arguing by 

 analogy, feasible to perform this operation by a method some- 

 what resembling that just described. Supposing, for example, 



■ By G. M. Whipple, E.Sc, F.R.Met.Soc, F.R. AS.. Superintendent of the 

 Kew Observatory. Kictimond (from the Qiin'lerly Jontna.1 C the Metcaro- 

 hgicnl .SKciety, vrl. U. No. 4S). 



