Comparisons of Mercury and Platinum Thermometers. 11 



the theoretical curves of best winding with a total of 530 

 turns. The coils were covered with gold-leaf, and in this 

 way connected with the surrounding metallic case to avoid 

 electrostatic disturbances. The resistance of each coil was 

 about 13 ohms. The sensibility of this galvanometer with all 

 the coils in parallel (Res. = 3*2 ohms) when tested with a 

 Wiess suspension system (two vertical magnets 27 mm. long, 

 1*5 mm. apart, made of small magnetized sewing-needles 

 and mounted on a thin strip of mica ; total weight about 

 45 milligrammes) was C = 6 x 10 -10 ampere; i. e., the current 

 required to produce 1 mm. deflexion on a scale 1 metre 

 distant when the period of a complete swing was 10 seconds. 

 When tested with a very light suspension system weighing 

 about 5 milligrammes (built of 10 or 12 magnets about 1 mm. 

 long each, made of fine watch-spring, tempered and strongly 

 magnetized, total weight about 6 milligrammes), the sensi- 

 bility under the same conditions was = 2*5 xlO -10 . All 

 the magnets used in the suspension systems were repeatedly 

 magnetized and boiled for several hours, so that magnets of 

 great permanency were obtained. Throughout the present 

 investigation the Wiess suspension system was used on 

 account of its greater freedom from external disturbances. 



Barometer and Standard Metre. — The barometer was of 

 the Regnault standard type with a tube about 1 m. in height 

 and 25 mm. internal diameter. The lower meniscus was 

 defined by a steel screw of known length. The barometric 

 height was measured on the standard metre bar placed beside 

 the mercury column by means of a cathetometer. The baro- 

 metric heights as measured by the scale on the cathetometer 

 column, merely as a check, were always found to be in fair 

 agreement with the values found by the standard metre. 

 This barometer was compared with a Fortin standard baro- 

 meter, and the agreement (to T \,- mm.) was as close as we 

 could read on the latter instrument. 



The standard Bartel and Dietrich's metre was divided into 

 millimetres, with the rulings on a strip of silver, imbedded 

 in a bronze bar of square section about 21 mm. on the side. 

 The corrections to the divisions of this metre were known 

 from previous comparisons with a standard steel metre which 

 had been compared with the Coast Survey and other standards 

 by Prof. Rogers. 



Standardization of Resistance-Box . 

 The resistance-box, mounted as already described, was 

 twice standardized with the greatest care. While we cannot 

 give within the limits of the present paper the details of the 



