12 Transactions of the South African Philosophical Society. 
It is curious that the monthly differences in both December, 1894 
and 1895, should be less than the November before and the January 
after, and in all probability the same thing would show in the mean 
results had there been no change of exposure at Kenilworth. 
The numbers in the table would appear to indicate that the Kim- 
berley days are warmer in summer and cooler in winter than those 
of Kenilworth. The great number of trees about Kenilworth may, 
perhaps, contribute something to this result; but it seems more 
likely to be due to an inherent defect in the Glaisher stand, however 
it may be modified : z.e., to the utter lack of protection which it can 
afford against the effects of radiation from the ground in such a 
climate as ours. How effective radiation may prove as a source of 
error in temperature observations, unless properly guarded against, 
will be seen from Table III., wherein a comparison is made between 
the approximate maximum surface-soil temperature at a depth of 
1 inch and the maximum temperature of the air. 
TABLE III. 
Mran Montuty Maximum TEMPERATURE OF THE AIR AT KENIL- 
WORTH COMPARED WITH THE MEAN SURFACE-SOIL TEMPERA- 
TURE AT 2 P.M.* 
Soil Temperature Monthly Maximum 
at 1 inch beneath Temperature of the Differences. 
Surface. Air. . 
Mean at Greatest 
1897. XIV. observed Mean Absolute 
Ss Ss’ T ae S-T S’-T’ 
January .. O87. 8) lass 84°9 94°7 So 13i38} ES) 11 
February ais AMO GS 91°3 99°8 +19°4 +18:0 
Miameln ast: Sica 112°8 83° 92°8 +14:0 + 20:0 
Aprile Sine 95:0 104°9 83°4 92°2 +11°6 +127 
WWE ale te 76°2 86°2 72°5 SLs) + 3°7 + 4°7 
Chae seals 63°3 69°0 65°2 T2°7 — 19 — 3°7 
rule alee 66°7 76°5 67°6 shor — 0°99 + 1°4 
August .. 80:9 | 88°2 73°0 83°3 + 6:1 + 49 
September . 92:0 104°8 Those lb 91:2 +13°9 +13°6 
October .. 104°2 119°3 85°3 96°6 +18°9 +22°7 
November | 110°3 LOST, 86°4 100°6 + 93°9 S808 )211 
December | 112°4 126°8 Ler Sige +20:°7 +29°3 
Means .. 923° 80°2° +11°9° 
Extremes . 2 GEse 100°6° + 26:2° 
* In October, 1898, it was ascertained that the depth of the bulb of the ther- 
mometer with which these observations were made had increased from 1 inch ° 
to nearly 2 inches beneath the surface. This may be accounted for partly by 
an actual sinking of the thermometer, and partly by a possible accumulation of 
drift sand upon the site. 
