302 
DR. C. CHREE: ATMOSPHERIC ELECTRIC POTENTIAL RESULTS AT KEW 
graphically in fig. 1 by the continuous line curve; the ordinates represent the 
mean values of P for the individual months, expressed as fractions of their arithmetic 
mean from the 12 months. 
Table I.—Potential Gradient (Volts per Metre). 
1898. 
1899. 
1900. 
1901. 
1902. 
1903. 
1904. 
Mean. 
January. 
174 
285 
189 
238 
164 
162 
194 
201 
February . . 
244 
298 
182 
247 
232 
174 
190 
224 
March . . 
216 
203 
180 
141 
152 
191 
177 
180 
April.... 
109 
139 
107 
140 
149 
165 
160 
138 
May .... 
97 
151 
96 
127 
93 
167 
130 
123 
June .... 
104 
105 
73 
114 
115 
156 
106 
111 
July .... 
104 
90 
86 
95 
86 
105 
117 
98 
August . . . 
125 
127 
104 
87 
111 
129 
118 
114 
September . . 
163 
83 
118 
117 
104 
135 
130 
121 
October . . . 
161 
175 
133 
191 
123 
141 
147 
153 
, November . . 
250 
167 
157 
215 
178 
172 
259 
200 
December . . 
184 
328 
263 
162 
237 
250 
273 
243 
Mean . . 
161 
179 
141 
156 
145 
162 
167 
159 
Table I. shows at once how variable an element P is. It is at least doubtful 
whether even seven years are sufficient to give an annual inequality which can be 
relied on as fully representative in its details. There is unquestionably a well-marked 
minimum near midsummer. There seems also to be a second minimum in January, 
but whether this would be shown—or if shown would be as prominent—if we had 
70 years instead of 7 is open to doubt. The fact that the largest mean monthly 
value did not occur in January in any one of the seven years, whilst occurring 
four times in December, is so far confirmatory. In November, December, and 
January, however, fog is a very uncertain while very important influence. During 
thick fog in winter P is usually abnormally high, 400 to 600 volts—or practically 
double the ordinary gradient—being not infrequently met with. Thus in winter the 
mean value of P varies much from day to day even in fair weather. For instance, in 
December, 1898, the last three of the selected days gave a mean value for P of 
380 volts, whilst the corresponding mean from the first seven days only slightly 
exceeded 100 volts. So, again, in January, 1899, the mean values of P from the 
first five and the last five of the selected days were roughly 160 and 400 volts 
respectively. The figures in Table I. itself, e.g ., those for December in the years 
1899 and 1901, afford examples almost as striking. 
The mean value for the year appears lower for the three consecutive years 1900 to 
1902—-which cover the period of sun-spot minimum—than for the others. As this 
phenomenon also appears when we confine ourselves to the six summer months, April 
