— 223 — 
eclipse-effect, this effect is an increase before and a decrease 
after totality, but it is difficult to say how much, if any, 
of the departures shown are attributable to the eclipse. 
The Z-curve shows a strong depression during eclipse-time, 
but here again it is uncertain whether this decrease is to 
be associated with the eclipse. 
Total Eclipse of the Sun of August 19, 1887. 
Course of central line: Russia, Siberia, Japan, Minimum 
distance from Zi-Ka- Wei + 15° N., maximum phase + 1* 55” p. 
The three curves for Zi-Ka- Wei are quiet, only the D-curve 
shows some irregularity, the needle moving from 12» 46" —51" 
for 0/15 to the West, discontinuating the easterly movement; 
but the next day the same irregularity, only a bit smaller, 
occurs at 12^ 40", 
I have plotted the D-curve with the aid of readings of 
the eurve taken by myself together with the published hourly 
readings. The average hourly values of the month (the year 
1887 being magnetically quiet) have been used for the 
construction of a normal curve. 
The ordinary daily variation on the eclipse-day is larger 
than the average one for the month; the maximum is reached 
at noon, but this is nothing particular. 
After the irregularity at 19^ 46"— 51", a slight wave is to 
be observed from 1^ p till 3^ p; which may be looked at as 
an eclipse-effeet. Again this effect is a deflection increasing 
and deereasing with the obscuration, ranging over 0.5. 
The H-curve of the magnetogram looks fairly regular and 
of a smooth course, but the hourly values for that day, 
published in the Bulletin are not regular. Perhaps these 
irregularities are the consequenees of a defective correction 
for temperature. 
- According to information received from the Rev. DE Mowrey 
the change for 1° C amounted to 48 y and 49 7 resp. for H 
and Z. The readings of the magnetograph-thermometers were: 
