NA TURE 



\7an. 8, 1874 



above and slightly behind the house. The solar rays falling on 

 the objects in the court-yard were transmitted through the 

 shutter holes, There being no other light in the room, and the 

 rays being strongly scattered by the rough whitewashed wall, the 

 rays were sufificiently powerful to produce an image on the retina 

 of an observer in whatever part of the room he might be ; the 

 room became, as it were, the box of a large camera. 



On intercepting the rays with a smooth oval looking-glass, 

 they were not, of course, scattered, and no image was visible on 

 the glass, but the image could be reflected from the looking-glass 

 to any part of the wall which contained the shutters through 

 which the rays passed. The appearance produced when a ser- 

 vant was made to stand in the required position, was singular. 

 A full-length (inverted) coloured figure appeared in an oval 

 frame of bright white light, much larger, of course, than the 

 looking.glass. The white light was produced by the glare from 

 the gravelled yard, shadows on which were reproduced. 



A dog-cart and horse were imaged on the wall most cjearly, 

 the chesnut colour of the hore being very distinct. The whole 

 phenomenon was always producible at any time when the sun 

 was in the proper position above the house. 



Are not mirages of one class, i.e. the appearance of inverted 

 images in clouds, produced in a similar way ? The rays from a 

 figure might pass through an opening in one cloud to the face of 

 another otherwise unilluminated, and be thence scattered. I 

 believe I have seen this explanation given somewhere, but I 

 cannot remember where. 



E. C. Buck 



N.W. P., India 



ON TEMPERATURE CYCLES* 



O INCE the discovery of an eleven years' period in thepbe- 

 »-' nomena of solar spots, several corresponding periods 

 (it is now well known) have been demonstrated in terres- 

 trial phenomena, more especially in those of magnetism, 

 auroras, cyclones, and rainfall. With regard to weather 

 changes, it has been thought by Dove, that the tracking 

 of a cycle in these could not, theoretically, be made an 

 object of research ; and that while some indications of a 

 periodicity might appear, a great part of the complicated 

 changes named must be, from the nature of the case, 

 quite unperiodical. The series of observations by Dove 

 on the subject led him to the conclusion (l) that diver- 

 gences from the normal, especially those of temperature, 

 are not local, but spread over large surfaces ; but (2) that 

 negative divergences, in one region of the earth, are com- 

 pensated by positive in another ; and conversely. That 

 the compensation is perfect, and that the quantity of heat 

 annually given by the sun is constant, has been affirmed 

 also by Maury and others. 



The data on whi :h this conclusion is based are limited. 

 They appeared quite insufficient to a German physicist. 

 Dr. W. Kdppen, who has recently been led to undertake a 

 wider investigation of the subject. He has communi- 

 cated to the Austrian Society for Meteorology a prelimi- 

 nary notice of his inquiries and results (Zeilscluiff, Aug. 

 and Sept. 1873), which will be found of considerable 

 value. 



We may first note here his materials and method. He 

 furnishes a long list of places from which observations 

 (more or less extensive) have been h id ; and in his first 

 table he givts the divergences of temperature of individual 

 years (1S20 — 71) from the average temperature, and for 

 the following regions : India, Tropical America, Temperate 

 South America, South Africa, Australia, China, and Japan, 

 Mediterranean region, Southern United States, Western 

 U.S ,WesternCentralEurope,Austria,South Russia, South- 

 west Siberia, East Siberia,Central part of U.S., Atlantic 

 Stat s, British Islands, North Germany and Netherlands, 

 North- West Russia, North-East Russia and Ural, North- 

 West America, North East America, Iceland, Northern 



♦ Abstract of paper by Dr. W. Kopptn 

 MetcorohgU. 



llie Austrian Zcitschri/t fur 



part of Europe. [The particular towns, &c., arc given, 

 and the author's purpose partly is, that the list may be 

 supplemented by other series of observations (which he 

 has not been able to see), being sent to the Central Phy- 

 sical Observatory at St. Petersburg, where he has chiefly 

 been prosecuting this research.] The periods of obser- 

 vation ranged from three to thirty years ; the average was 

 taken from several years' observations. In many obser- 

 vation-series, the yearly average had to be calculated for 

 the first time. Series of six years' length v/ere the shortest 

 admitted, and such short series only by way of completing 

 the longer. The original sources of Prof. Dove's material 

 were consulted. 



A second table shows the divergences of temperature 

 in various regions for the years 1768 — 1819. By way of 

 condensing, a third table is given, in which the material 

 from 1820 — 71 is arranged in five series, one of which 

 represents the tropics, and the four others four successive 

 ex-tropical zones. The zones are not bounded by deter- 

 minate parallels of latitude, but it was sought to combine 

 approximately equal material of observation and earth 

 surface. 



On comparison of the curves of Table III. with the sun- 

 spot curves (according to Wolf), a striking correspondence 

 at once appears, as far as the year 1854. In the tropics, 

 the maximum of heat occurs \ — 1\ years before the spot- 

 minimum ; in the ex-tropical zones, on the other hand, it 

 occurs after the minimum ; in some cases (in the forties, 

 e.g.) as much as three years after. The regularity and 

 extent of the variations diminish from the tropics to the 

 poles. 



It is further noticeable that as the interval from maxi- 

 mum to minimum of the spots is always greater than 

 that from minimum to maximum, a corresponding in- 

 equality occurs in the temperature changes. 



On these results Dr. Koppcn remarks that, while there 

 is evidently some connection between the two kinds of 

 phenomena, the sun-spots do not act directly by darken- 

 ing a part of the solar disc ; for, as the temperature of 

 the earth's surface is a function of the solar radiation, 

 the change in the former must follow that in the latter ; 

 but the opposite occurs, as we have seen, in the tropics. 

 It is probable that the temperature of the sun's surface is 

 (from some unknown cause), at its highest one or two 

 years before the minimum of the sun-spots. That the 

 spots (if we suppose them to be solid bodies) take so 

 long to melt that their minimum only occurs after the 

 maximum temperature of the earth's surface, is not 

 remarkable, considering their size. 



If we consider the period 1800 — 71, we find a section 

 of about 40 years, with marked periodic variation, 1815 — 

 54, and two periods, before and after, showing great dis- 

 turbances, (say) 1792 — 1815, and 1854 — 66. Whether in 

 1865 we have again entered (as the curve would seem to 

 indicate) on a time of distinct periodic variation, will 

 doubtless appear in the next ten years. 



The observations before 1800, again, show such anoma- 

 lies in the temperature, that we should almost doubt the 

 existence ot connection with the sun-spots were it not 

 for the convincing evidence of the years 1S15 — 54. 

 We find all possible cases, from complete indirle- 

 rence of the temperature in contemporaneous change 

 of the sun-spots (1750 — 71), and a short correspon- 

 dence of both (1772 — 77), to a well-marked and regu- 

 lar variation of temperature (1777 — 90), which stands 

 to the sun-spot curve, in exactly the opposite relation to 

 that found in 1816 — 54. True, the observations here are 

 only from a small fraction of the earth ( West Europe and 

 the New England States) ; but the. continuance of the 

 same curve shows the normal variation in 1S16 — 54 quite 

 distinctly. The estimation of the spots previous to 1826 

 is somewhat arbitrary, but an error such as that the maxi- 

 mum is put in the place of the minimum cannot be sup- 

 posed. And lastly, if it be urged that the turning points 



