456 



SCIENCE. 



[N. S. Vol. XXIII. No. I 



Using a stiff card, the sphere is projected 

 on the plane of the meridian. The horizon 

 and twilight circles eighteen degrees below 

 are also drawn. Upon another smaller 

 circular card, pivoted to the first at its 

 center, the six-o'clock hour-circles through 

 the poles, the equator, and diurnal paths 

 of the sun at the solstices are projected in 

 straight lines. Harmonic projections of 

 the hour-circles for every twenty minutes 

 are marked on these. 



By revolving the movable card the pole 

 can be set to the proper elevation for any 

 latitude, and the number of hours of dark- 

 ness, twilight and daylight can be read. 

 The time of sunrise and sunset and the 

 meridian altitude of the sun at noon or 

 midnight can also be measured. 



These fundamental phenomena should be 

 made clear even in preparatory schools, but 

 from our observation of college students it 

 is evident they are not. To put such a 

 piece of apparatus into the hands of each 

 student, with which he can solve a set of 

 problems, would be another move in the 

 direction of individual laboratory work in 

 general astronomy which it seems so di,ffi- 

 cult to seciire. 



The model has been entrusted to the 

 Arthur Hall Scientific Company, of Bos- 

 ton, for reproduction, in the hope that it 

 may be of service. 



300) of the Y scale. The corrections are 

 given to ten-thousandths of a millimeter 

 and are accurate to within a few units of 

 the last place. 



The irregularities of the X micrometer 

 throughout three revolutions have been in- 

 vestigated and a table of the corresponding 

 corrections determined by the method out- 

 lined in Briinnow 's ' Spherical Astronomy, ' 

 p. 426, fourth German edition. Two inde- 

 pendent determinations yield results differ- 

 ing by about .0003 mm. in the maximum. 



The straightness of the cylinder which 

 determines the Y axis and of the bar which 

 determines the X axis was also tested, as 

 well as the perpendicularity of the direc- 

 tions of these two parts of the apparatus. 

 The cylinder and bar were found to be very 

 satisfactory and their directions are made 

 perpendicular by adjustment. 



The measurement of the distance be- 

 tween two well-defined points on a plate, 

 first as a difference of X coordinates and 

 then as a difference of Y coordinates, yields 

 results differing by slightly more than the 

 error of measurement, indicating that the 

 Y scale is not exactly parallel to the plate. 

 The maladjustment in this particular 

 would introduce an error into the measures 

 that would in general be entirely negligible. 



Tables are added for applying the neces- 

 sary corrections to measured coordinates. 



The Eepsold Measuring Apparatus of the 

 Students' Otservatory, Berkeley, Cal.: 

 Burt L. Newkirk. 



This paper contains an investigation of 

 the instrument, with a determination of 

 corrections which must be applied to meas- 

 ures made with it. The method employed 

 in measuring division errors of the scales 

 is essentially the one employed by Gill for 

 the scales of the Cape heliometer and Rep- 

 sold measuring apparatus. Corrections 

 are determined for each line of the X scale 

 and for seventy lines ( from line 230 to line 



The Forty-foot Camera of the U. S. Naval 



Observatory Eclipse Station at Guelma, 



Africa: W. W. Dinwiddie. 



This instrument was mounted horizon- 

 tally and the light was led to it by the mir- 

 ror of a Gaertner coelostat. Three ven- 

 tilating doors were made in the lower side 

 of the forty-foot tube, which, operated in 

 connection with the trap doors in the roof 

 of the building at the plate end of the tube, 

 gave perfect ventilation and kept the whole 

 instrument cool. 



The shutter used in making the exposures 



