DEFLECTION OF LIGHT DYSON AND OTHERS. 145 
Thermometer readings, July 10, 74.4°; July 13, 73.7°; July 14, 
71.9° ; July 16, 72.3° ; July 17, 72.3°. 
By July 18 we had obtained a sufficient number of reference photo- 
graphs. Dismantling of the instruments was commenced, and the 
packing was completed on July 21. We left Sobral on July 22, 
leaving the packing cases in the hands of Messrs. Nicolau and 
Carneiro to be forwarded at the earliest opportunity, and arrived 
at Greenwicli on August 25. 
The observers wish to record their obligations to Mr. Charles Booth 
and the officers of the Booth Line for facilitating their journeys to 
and from their station at a difficult time. 
PHOTOGRAPHS TAKEN WITH THE 4 -INCH OBJECT GLASS. 
16. These photographs were taken on 10 by 8 inch plates. By 
suitably mounting the camera it was made possible to obtain seven 
stars on the photographs, viz, Nos. 2, 3, 4, 5, 6, 10, and 11 of the table 
in section 5. Of the eight photographs taken during the eclipse 
seven gave measurable images of these stars, the other plate (No. 6) 
taken through cloud only showing a picture of the prominences. 
Plates of the same field taken under nearly similar conditions as 
regards altitude were taken on Jnly 14, 15, 17, and 18 (civil date). 
Of these photographs, the second taken on July 14 with two ex- 
posures (referred to as 142a and 1426), two photographs taken on 
July 15 (referred to as 15^ and I52), two on July 17 (17^ and I72), 
and the second photograph on July 18 (18,) were measured for com- 
parison with the eclipse plates. 
17. The micrometer at the Eoyal Observatory is not suitable for 
the direct comparison of plates of this size. It was therefore decided 
to measure each plate by placing, film to film upon it, another pho- 
tograph of the same region reversed by being taken through the 
glass. A photograph for this purpose was taken on July 18. This 
plate is regarded merely as an intermediary between the eclipse 
plates and comparison plates and is referred to as the scale plate, 
being used simply as a scale providing points of reference. In all 
cases measurement was made through the glass of the scale plate, 
adjusted on the eclij3se or comparison plate which was being meas- 
ured, so that the separation of the images on the two plates did not 
exceed one-third of a millimeter. The plates were held together by 
clips which insured contact over the whole surface. This method 
of measurement was found to be very convenient. Each plate was 
measured in two positions, being reversed through 180°, and the 
accordance of the result showed that the method of measurement was 
entirely satisfactory. 
