SECT. 1] GRAVITY AT SEA 151 



of the assumptions about the mass between the gravity meter and the sea-level 

 surface that are made for measurements on land. The free-air anomaly is ex- 

 pressed in milligals (1 mgal = 0.001 cm/sec-). 



Free-air anomalies are caused by differences between the distribution of 

 masses on the real earth and on the earth model assumed in the International 

 Formula. The most rapid variations of free-air anomaly, though usually not the 

 most important on a large scale, are due to the changing effect of the bottom 

 topography. Small-scale variations in the free-air anomaly are often closely 

 correlated with changes in the depth of water, as is plain from Figs. 11, 12 and 

 20, and these changes must be removed or allowed for before the other reasons 

 for variation, such as geological structures beneath the sea floor, can be 

 investigated. 



The depth of water in the immediate vicinity of the ship is the most in- 

 fluential in affecting the gravity anomaly, and the effect of a given topographic 

 feature decreases rapidly with increasing distance. As the ship moves into 

 shallower water, there is an increase in the mass immediately beneath it and a 

 consequent increase in free-air anomaly. One method for removing this direct 

 correlation with topography is to compute the increase in gravity at the j)oint 

 of observation that would be caused by substituting rock of a certain density 

 for the sea-water. This "topographic correction" is added to the free-air 

 anomaly and the new anomaly so produced is called the Bouguer anomaly. 

 The gravity profile will show no correlation with bottom topography if the 

 correct density is chosen for the added rock and there are no concealed masses. 

 Tills is illustrated in Fig. 11, which shows j)rofiles of the free-air anomaly and 

 Bouguer anomalies computed for densities of 2.1, 2.3^ 2.5 and 2.67 g/cm^ across 

 Seamount Jasper (Harrison and Brisbin, 1959). The free-air profile shows a 

 sharp maxinium over the seamount ; this maximum is not quite removed when 

 a density of 2.1 is adopted. The profile for 2.3 g/cm^ density shows no indication 

 of the presence of the seamount while those profiles for higher assumed densities 

 show a minimum over the peak. Replacing the sea-water by rock of 2.3 g/cm^ 

 density eliminates the gravitational effect of the seamount altogether — it is 

 (hypothetically) buried in material of its own density and is no longer detectable 

 gra vitationally . 



The computation of Bouguer anomalies at sea involves the hypothetical 

 addition of mass to the ocean and so these anomalies do not give a true picture 

 of the variation of mass per unit area over the earth's surface. They are always 

 large and positive over the deep sea, because of the added mass, and large and 

 negative over elevated land areas, because here mass is subtracted in the com- 

 putation. Fig. 11 shows the increase in size of the Bouguer anomaly with 

 increase of the assumed density ; although a density of 2.3 g/cm^ is the correct 

 one to use in this profile, because it is the one that best eliminates the gravita- 

 tional effect of the seamount, it certainly is not the correct one to use for 

 comparing the mean level of anomaly in the Pacific with the level in shallow 

 water close to the shore. In the latter case it is proper to use the mean density 

 of the continental block between sea-level and the level of the sea floor in the 



