86 TlEPORTS ON THE STATE OP SCIENCE. — 1919. 



developed on board ship ; on the Morea this was done in the refrigeratot, 

 and the results are better than thosr- obtained on the Ascauius when 

 developing was carried out in a cabin. Rodinol and Eastman's cine- 

 matograph films were very satisfactory. During the voyages over 5,000 

 readings of the thermometers, &c., were made and 100 feet of film 

 developed. 



Reduction of Results. 



A measuring machine originally constructed by Hilger for spectroscopic 

 work and accurate to -002 mm. was used. Each film was placed so that 

 its length was perpendicular to the direction of travel of the micrometer 

 eye-piece. A double thread was sighted along the line of hollows, then 

 along the line of crests, then upon the images of the engraved lines, and 

 finally upon the top edge of the exposed portion of the film. This was 

 done in three different parts of each exposed strip, the results tabulated 

 and the distances calculated. 



Two errors peculiar to photographic registration have to be corrected :— 



(1) The scale is engraved on the glass, and unless the ray of light 

 passes horizontally over the curved meniscus the relative positions of 

 mercury and scale will not be correctly shown upon the film. Assuming 

 that the shape of the meniscus is known, the correction can be calculated 

 for different heights of the mercury above the horizontal ray, but it is 

 more satisfactory to find it experimentally by taking a number of photo- 

 graphs in harbour when the barometer is steady, thus: — the whole 

 instrument is moved up and down, so that the mercury surface is at 

 various heights above and below the horizontal ray, and the distances 

 on the film of the mercury surface from one of the engraved lines are 

 tabulated against the depth of the mercury from the top of the wmdow 

 through which the light passes, a distance easily calculated from measure- 

 ments of the widths of the exposed band. To correct for any variations 

 in the barometer during the test it is advisable to return to a fixed reading 

 at frequent intervals. Experiments according to this programme were 

 carried out in Cape Town Harbour on July 13, 1914. 



The correction to the observed reading of one of Hecker's barometers 

 is approximately -008 mm. for a difference of 0-1 mm. in the width of 

 the exposed strip. 



(2) The brightness of the lamps, the rate of running of the film, and 

 the time of development being variable, all developed images are not of 

 equal blackness. The position of the boundary of this image depends 

 to some extent upon its density, hence it is necessary to introduce a correc- 

 tion on this account. 



This can be evaluated by taking a series of records with constant 

 barometer with illuminations of different powers. The density of each 

 deposit may be measured by a density meter, and the corresponding 

 measurement of the film plotted against it. On Hartmann's scale of 

 densities the correction varies from - 0-02 mm. for density 5 to + 0-02 mm,, 

 for densitv 75. The densities of the films varied over very narrow limits, 

 so for this reason and others which will appear, these corrections were 

 regarded as too small to merit application to any of the figures obtained. 



