October 28, 1910] 



SCIENCE 



607 



ial barometric height referred to normal grav- 

 ity, the outstanding difference being the meas- 

 ure of the gravity anomalies within the inevit- 

 able observational errors. He had made a 

 cruise on a passenger steamer in 1901 from 

 Hamburg to Eio de Janeiro and back to Lis- 

 bon; again in 1904 he made further cruises in 

 the Indian and Pacific oceans and in 1909 

 also in the Black Sea. As the result of Heck- 

 er's experiences, Professor Helmert did not 

 think it possible to get anything of value on 

 such a small vessel as the Galilee, and we ac- 

 cordingly made no attempt. 



However, on the Carnegie, it was decided to 

 make as frequently as possible determinations 

 of the temperature of the boiling point of 

 water with the prime view of obtaining the 

 data required for controlling the corrections 

 of our' aneroids. The instrumental equipment 

 was in accordance with this chief purpose, and 

 hence only two boiling point apparatuses, fur- 

 nished each with a thermometer, and an ordi- 

 nary Green marine mercurial barometer, were 

 provided. In all one hundred and two de- 

 terminations were made, representing seventy- 

 five different points, four of which were the 

 ports Brooklyn, Falmouth, Madeira and Ber- 

 muda. It should be said that the observer. Dr. 

 C. C. Craft, had no idea of the possible use of 

 his results for gravity ; however, a very search- 

 ing examination has convinced me that with 

 the necessary refinement in instrumental 

 equipment and in method of observation it 

 wiU be possible to obtain gravity results on 

 the Carnegie worth while. While the results 

 on the first cruise can not be used for deter- 

 mining the gravity anomaly over any re- 

 stricted area, a general deduction can be 

 drawn, in view of the large number of obser- 

 vations and the varying conditions under 

 which they were made, which harmonizes with 

 the general conclusion of Hecker's work, 

 namely, that gravity is in general normal 

 over the deep oceans, the defect in the density 

 of the aqueous material above the ocean bot- 

 tom being made up very nearly by increased 

 density of the material below the ocean bot- 

 tom. If we average our results then the mean 

 anomaly over the deep part of the Atlantic 



Ocean, three to seven kilometers, differs from 

 the mean over the shallow part along the 

 coasts, five to two hundred meters, by about 

 -|- 0.04 of a dyne, meaning that there is a 

 slight defect in gravitational force over the 

 deep ocean by that amount as compared with 

 the force over shallow water. This is to be 

 regarded merely as a provisional result, the 

 indication being that the final reduction may 

 diminish the quantity to -f 0.03 or even to 

 + 0.02 dyne. Our result is in the same direc- 

 tion as Hecker's conclusions, he getting fig- 

 ures on the order of -\- 0.02 to + 0.055 of a 

 dyne. A difference of .035 of a dyne would 

 correspond approximately to an error of 

 0.001 of a degree in the temperature of the 

 boiling point of water. The average difference 

 in the temperatures of the boiling point for 

 the two thermometers used and for a single 

 determination was 0.003 of a degree; the 

 average result above depends upon 75 sta- 

 tions. In his latest work Heeker employed 

 nine specially constructed mercurial barom- 

 eters and six thermometers and six boiling 

 point apparatuses. 



In connection with this investigation I have 

 had occasion to examine into the various 

 tables for obtaining the atmospheric pressure 

 corresponding to the temperature of the boil- 

 ing point of water, the latest of these tables 

 in general use being those of Wiebe's given in 

 Landolt-Bornstein's " Physikalisch-Chemische 

 Tabellen." The most recent observations ap- 

 pear to be those of Holborn and Henning. 

 For the purposes of gravity work, it is essen- 

 tial to be able to obtain accurately the atmos- 

 pheric pressure for a comparatively limited 

 range extending below and above 100° C; the 

 observations on which the tables are based on 

 observations at larger intervals and the inter- 

 polation is accordingly somewhat uncertain. 

 It is quite possible that the atmospheric pres- 

 sure as talvcn from the tables may be out .05 

 to 0.1 mm. which corresponds to 0.065 to 0.135 

 of a degree in gravity. When dealing with 

 only differential results of gravity, as we are 

 in the present instance, the tabular errors are 

 somewhat eliminated, though not wholly. I 

 desire to bring this problem of more accurate 



