December 1, 1905.] 



SCIENCE. 



705 



curacy, and that if it were necessary to still 

 further increase the accuracy of the results, 

 this could be done by spending additional time 

 in the observations. When observations can 

 not be secured on a complete ' swing-,' but 

 simply on the regular' course of the ship, it 

 is not always possible to mathematically con- 

 trol the deviation corrections applied, owing 

 to accidental conditions entering in. While 

 these corrections in the case of the vessel 

 employed are comparatively small as compared 

 with those of other expeditions, they are of 

 sufficient amount to require being taken into 

 account in securing data of the precision 

 requisite for the solution of some of the 

 greater problems referred to above. Consid- 

 erable time would be saved were it possible to 

 have a vessel entirely non-magnetic so that 

 the question as to corrections to be applied on 

 account of magnetism of any portions of the 

 ship need not be considered. 



The results thus far secured by the Galilee 

 on her cruises from San Francisco to San 

 Diego and from there to the Hawaiian Islands, 

 as well as some results obtained by the Coast 

 Survey vessel in the Pacific Ocean — the Pat- 

 terson — ^proved that the latest magnetic charts 

 are systematically in error, as far as the mag- 

 netic declination is concerned, to the extent of 

 from one to two degrees, the charts giving too 

 low values of easterly declination. The lines 

 of equal dip appear to be correct on the av- 

 erage within about one third of a degree. The 

 lines of equal horizontal intensity are sys- 

 tematically erroneous to the extent of one 

 twentieth to one thirtieth part of the absolute 

 value — fully ten times the error of the ob- 

 servation — the charts giving too high values. 

 A consideration of the values obtained by the 

 Coast Survey vessels in the Atlantic Ocean, 

 especially between Baltimore and Porto Rico, 

 likewise shows that the intensity charts give 

 values too great by about the same ratio as in 

 the case of that portion of the North Pacific 

 Ocean considered above. 



The president followed with some extended 

 remarks on the subject. 



The 606th meeting was held November 4, 

 1905. 



Mr. H. B. Brooks described, by invitation, 

 ' An Efficiency Meter for Incandescent Lamps' 

 developed by Mr. Hyde and himself for use at 

 the Bureau of Standards. The purpose of 

 this is to give a direct reading of the quotient 

 of the watts used by the candle-power; for 

 commercial 16 c. p. lamps this quotient is 

 three to four. A Weston wattmeter is used, 

 but enough extra resistance is added in the 

 shunt circuit (which has normally some 2,000 

 ohms resistance) to bring the deflection of the 

 needle down to ten times the quotient, as from 

 64 to 40. Since this resistance must vary with 

 the observed candle-power, part of it is wound 

 on a block of carefully calculated form and a 

 contact piece carried along with the photo- 

 meter screen cuts out resistance as the candle- 

 power increases. The instrument is reliable 

 to about one per cent. 



Mr. W. P. White then presented ' A Thermal 

 Study of the Mineral WoUastonite ' made by 

 himself and Messrs. Allen and Wright. This 

 substance, often found in lavas, was studied 

 to get a probable limit to the temperature 

 that the lava had reached. 



Calcium metasilicate, CaSiOj, exists in two 

 forms; one, wollastonite, stable below about 

 1,180°, and a monoclinic form stable above 

 that temperature. If wollastonite is heated, 

 it changes readily to the other form, but to 

 get the reverse change is often a matter of 

 some difficulty. Hence, on cooling a charge 

 of the melted material, if crystallization oc- 

 curs, as it usually does, above 1,180°, pseudo- 

 wollastonite is found. It is only now and 

 then that the undercooling is great enough to 

 allow wollastonite to crystallize. Wollastonite 

 can be formed readily, however, by chilling 

 melted material so that it becomes glassy at 

 ordinary temperatures, and then heating this 

 glass to a dull red heat. The melting point 

 of the pseudo- wollastonite is 1,512°. Nine 

 determinations on four separate samples 

 showed a maximum variation of 2-|° in the 

 determination of this point. 



In locating the inversion temperature in 

 the electric furnace, great help was obtained 

 by the use of control elements which gave 

 simply the furnace temperature and enabled 



