2l8 



NA TURE 



[January 7, 1892 



representing the average behaviour of thermometers of 

 this class. He says : — 



" A few days' observations under identical conditions 

 are not sufficient to determine the correction with any 

 approach to accuracy. The thermometers are so variable 

 in their indications, that, in one ordinary case that I 

 worked out, it would seem that at least forty-four months' 

 comparative readings would be required to furnish an 

 averaj^e correction with a probable error of only one-tenth 

 of a degree. The differences between the indications of 

 two thermometers placed side by side are in very many 

 instances subject to an annual variation, showing that 

 the correction to a common standard cannot be made by 

 adding or subtracting a fixed quantity, but that the 

 amount of this correction is variable, and perhaps capable 

 of being expressed as a function of the temperature indi- 

 cated. The older instruments, even after correction, on 

 the whole give lower readings than the new ones. Some 

 of the latter, when compared with the oldest thermometers 

 of the set, appear to fall off considerably in sensitiveness 

 even in the short period of twelve months. . . . But 

 sometimes an instrument two or three years old decreases 

 in sensitiveness more rapidly than a perfectly new one ; 

 sometimes also an instrument, after remaining nearly 

 constant in its indications for several months, as 

 compared with the oldest of the set, suddenly shows a 

 rapid and unaccountable falling-off in sensitiveness." 



And he concludes : — 



" The indications of the instruments are thus in most 

 cases totally unreliable, and the observations compara- 

 tively worthless. The only possible exceptions I can see 

 to this sweeping condemnation are observations made 

 with instruments which have been in constant use for ten 

 years or more, and which may perhaps be assumed to 

 have arrived at a constant condition as regards sensitive- 

 ness." 



Since the average duration of a sun-thermometer, 

 under the conditions of Indian observatories, is only about 

 three years, it is obvious that instruments that have stood 

 the prescribed test can be but few. But it was with such 

 a thermometer that were obtained the valuable results 

 published by Mr. Hill in the Journal of the Asiatic Society 

 of Bengal in 1883 and 1886, which aflford the only direct 

 evidence yet on record of an eleven-year variation of the 

 solar heat. 



The duration of sunshine has now been recorded at 

 five Indian Observatories with the Stokes-Campbell sun- 

 shine recorder for periods of from four to seven years, 

 and the average results are given in Mr. Eliot's Report^ 

 together with those of the year 1889. The stations are 

 all in Northern India, one only, Calcutta, being within 

 the tropics. Allahabad and Lahore show the highest 

 proportion of sunshine, viz. 69 and 68 per cent, respec- 

 tively of the possible maximum. Jeypore has but little 

 less, viz. 65 per cent. Calcutta follows with 59 per cent., 

 and Dehra, at the foot of the Himalayas, shows the 

 lowest average, viz. 49 per cent. At St. Aubin's, in 

 Jersey, the sunniest station in the British Isles, the pro- 

 portion is 39 percent. In the absence of any record from 

 Southern India, it cannot be positively asserted that the 

 Indo-Gangetic plain is the sunniest portion of India, but 

 judging from the registers of cloud proportion, which are 

 regularly kept at all Indian stations, there can be but Uttle 

 doubt that such is really the case. 



Another kind of observations nearly related to the 

 above are those of the temperature of the ground. These 

 NO. 1 158, VOL. 45] 



have been made at the same five Observatories that have 

 furnished the sunshine records : at Calcutta since 1878, 

 and at the other stations during periods of from four to 

 ten years. One feature is common to all of them. In all 

 cases the mean temperature of the ground is some degrees 

 higher than that of the air, the excess varying, however, 

 considerably at different stations, and at the same station 

 in different years , as well as, of course, at different 

 seasons of the year. This general fact was observed 

 many years ago by the late J. Allen Broun at Trevandrum, 

 and has also been known for some years in the case of 

 Calcutta and Allahabad ; and it was remarked by the 

 late Prof. Hill that it is probably characteristic of hot 

 climates in general. He considered it probable that 

 there is 



" a difference in the opposite direction between the air 

 and ground temperatures in high latitudes ; for, owing to 

 the circulation of the atmosphere, and the constant mixing 

 together of its several parts, the air temperature must be 

 more uniform all over the earth than it would be were it 

 determined for each place solely by the balance between 

 insolation and loss of heat by radiation into space ; while 

 the temperature of the ground is more directly dependent 

 on the balance between the gain and loss of heat by 

 radiation." 



I am not aware that this interesting speculation has 

 hitherto been verified. 



According to the present registers, Jeypore (where the 

 ground consists of loose dry sand) shows the greatest 

 excess of ground surface temperature, viz. 6" '3 ^ on the 

 mean of the year ; and at Lahore (where the"'ground is a 

 sandy loam) it is as much as 5'''8. At Allahabad it appears 

 to be about 3°'8, and at Calcutta 2' 7. It is therefore in a 

 great measure dependent apparently on the dryness of 

 the climate, since the mean annual rainfall of these four 

 places is 25, 21 , 38, and 62 inches respectively, and the mean 

 relative humidities 50,50, 61, and ^'j percent, of saturation. 

 The ground temperatures here considered are those of 

 the surface. At Calcutta the temperature increases 

 rapidly with the depth, so that at 3 feet deep it is i'^'5 

 warmer than at the surface, and at 6 feet i°'6 warmer. 

 This is probably to be attributed to the decomposition of 

 organic matter, with which a bed of foetid quicksand at a 

 depth of 40 or 50 feet is highly charged, and which, when 

 freshly excavated, is distinctly warm to the touch.- 



At other stations there appear to be some remarkable 

 irregularities in the temperatures at different depths. Thus 

 at Allahabad the average warmth of the ground decreases 

 i°7 between the surface and the depth of i foot, and then 

 increases i'^'2 to a depth of 3 feet ; at Jeypore it decreases 

 3°"3 down to i foot, and then increases, but somewhat 

 irregularly, by a total amount of o°'6 to a depth of 20 feet ; 

 and at Lahore it decreases r'5 to a depth of i foot, and 

 increases again o''7 to 3 feet. In the case of Allahabad, the 

 temperatures of which were fully discussed in a memoir 

 by the late Prof. Hill,^ some of these minor irregularities 



I Most of the figures quoted in this paragraph differ from those given in 

 Mr. Eliot's Report. In the Report the comparison is made between the mean 

 ground temperature of a few years and that of the atmosphere deduced from 

 three or four times as many, and in some cases the conditions of the Ob- 

 servatory have been changed. The figures in the text are derived from a 

 comparison of the same years. 



^ This bed extends apparently everywhere beneath Calcutta, and is the 

 cause of great instability to the more ponderous edifices, of which the great 

 Imperial Museum affords a striking example. In fact, in a certain sense, 

 Calcutta may be said to be a floating city. _ , ^ . 



3 ' Indian Meteorological Memoirs," vol. iv., Part iii.. No. v., ' Onthe 

 Ground I'emperature Observations made at the Observatory, Allahabad. 



