7h a 30.8 C 
10 31. 5 
rog 32. 7 
4 33. 3 
7 33.2 
31. 
Accordingly large errors will have originated from a slight 
discrepancy between the temperature of the magnet with the 
thermometer-reading. Such a difference is very probable to 
have existed in case of notable daily changes as have occurred. 
Also it is to be remarked that the daily variation for the 
months of August of the two years of small solar activity 
1887 and 1901 differ strongly, which is without doubt -the 
consequence of defective temperature correction in 1887, as 
the temperature-change in the magnetograph-room in 1901 
was very small, whereas it was great in 1887. (Windows 
were opened during the night in 1887!) 
I have taken readings of the H-curve for every fifteen 
minutes between noon and 4" p and made them agree with 
the hourly values published in the Bulletin with the aid of 
linear changing corrections. 
Departures from the daily mean on August 19, 1887. 
= D H. D. H Z 
a 35 -ilay 09» Pop 85 — 1087 —057 
1.1 —164 0.1 15 —2.0 — 9.8 
B — 74 0.0 30 —1.7 98 
.9 —180 0.0 45 —1.4 — 94 
I 215 6.1 3 0 —1.3 — 90  —0.5 
sl — 94 15 —(0.8 — 8.2 
.9 —11.5 30 —0.6 — 5.7 
B 18:3 45 —0.1 — 25 
‚BI —01 .4 0.3 Lë —0.2 
.9 —13.9 5 1.3 — 25  —032 
.5: —13.9 6 0.9 — 6.5 | —0.3 
.0 :—12,3 7 0.1 — 49  —1.0 
8 0, 4 13.9 — 0.3 
