On the Registration of the Chemical Action of Daylight. 235 



mental error may be estimated. Curves exhibiting the graduation 

 of several strips are also given ; and from these the author concludes 

 that the determinations agree as well as can be expected from such 

 photometric experiments, the mean error between the positions 40 

 and 80 min. en the strip in one series of graduations not exceeding 

 1 per cent, of the measured intensity. To each fixed strip a Table 

 is attached, giving the intensity of the light which must act for 1 

 second upon the standard paper, in order to produce the tints at 

 each millimetre of the length of the strip. 



The methods of exposure and reading are next described. The 

 exposure of the paper is effected very simply by pasting pieces of 

 standard sensitive paper upon an insolation band, and inserting the 

 band into a thin metal slide having a small opening at the top and 

 furnished with a cover, which can be made instantly to open or close 

 the hole under which the sensitive paper is placed. When one ob- 

 servation has thus been made, and the time and duration of the inso- 

 lation noted, the remaining papers can be similarly exposed at any 

 required time ; and thus the determinations can be very easily car- 

 ried on at short intervals throughout the day. 



The reading-instrument consists of a small metallic drum, fur- 

 nished with a millimetre scale, and upon which the graduated strip 

 is fastened. The drum turns upon a horizontal axis, and the inso- 

 lation band, with the exposed papers upon it, is held against the 

 graduated strip, so that by moving the drum on its horizontal axis 

 the various shades of the strip are made to pass and repass each of 

 the papers on the insolation band, and the points of coincidence of 

 tint on the strip and on each of the exposed papers can be easily 

 ascertained by reading off with the monochromatic soda-flame. 



In the next section of the paper the author investigates the 

 accuracy and trustworthiness of the method. This is tested in the 

 first place by making simultaneous measurements of the chemical 

 action of daylight by the new method and by means of the pendulum 

 photometer, according to the mode described in the last memoir, 

 upon which the new method is founded. Duplicate determinations 

 of the varying chemical intensity thus made every half- hour on 

 four separate days give results which agree closely with each other, 

 as is seen by reference to the Tables and figures showing the curves 

 of daily chemical intensity which are given in the paper. Hence 

 the author concludes that the unavoidable experimental errors 

 arising from graduation, exposure, and reading are not of sufficient 

 magnitude materially to affect the accuracy of the measurement. 

 As a second test of the trustworthiness and availability of the method 

 for actual measurement, the author gives results of determinations 

 made with two instruments independently by two observers at the 

 same time, and on the same spot. The tabulated results thus ob- 

 tained serve as a fair sample of the accuracy with which the actual 

 measurement can be carried out ; and the curves given represent gra- 

 phically the results of these double observations. From the close 

 agreement of these curves, it is seen that the method is available for 

 practical measurement. 



In order to show that the method can be applied to the purposes 



E2 



