562 RFJP0KT8 ON THE STATE OF SCIENCE. — lOlTi. 



line, and the contour of the bed of the Arabian Sea and the Gulf of Aden 

 is followed closely. The floor of the Red Sea is nowhere deep, and there 

 is a defect of gravity which is very pronounced as a coral shoal is ap- 

 proached, very small in the centre of the sea, and again marked in the 

 neighbourhood of Suez. In the Canal a defect of gravity appears which is 

 not easily explicable if it is real. The Mediterranean shows an excess at 

 first, but a defect over the deepest part. The ajjproach to Malta is 

 characterised by a rise in the value of gravity (in conformity with the 

 known tendency of island stations), which increases before leaving the 

 shallow water south of Sicily and again when on the ridge south of Sar- 

 dinia. One may infer either that this ridge is of great density, which may 

 account for its capability of supporting Corsica-Minorca, or else that the 

 graph should have been dropped down on account of some vagary of the 

 aneroid, when the gravity and sea-floor contours would fit together rea- 

 sonably. In either case some tendency for gravity to increase as the bed 

 of the Mediterranean rises is apparent. The approach to the Straits of 

 Gibraltar occasions a pronounced fall in the value of gravity ; such has 

 previously been observed on the edge of a land mass, even though the 

 water is shallow, e.g., south of Colombo and Bombay and in the Red Sea. 



Finally, in the extension of the Atlantic Ocean known as the Bay of 

 Biscay there is an indication of a defect in gravity, but not as pronovinced 

 as in the Indian Ocean, where the depth is the same. 



In fig. 11 the depths are plotted against deviations from the normal 

 values of gravity. For shallow water there is little regularity, thougli a 

 gciaeral reduction below the normal value, perhaps corresponding to the 

 known defect of gravity at coastal stations, but beyond a certain depth a 

 diminution of gravity is associated with increasing depth. 



The results, if confirmed, will very seriously limit the application of 

 the isostatic theory of the earth's equilibrium, since over the Indian 

 Ocean the value of gravity is -2 to 3 cms. /sec- less than that demanded 

 by the mathematical expression of Pratt's hypothesis, a very appreciable 

 amount in gravitational units. The compensation appears to be less 

 complete than the simple theory had led us to hope. 



The above suggestions are put forward tentatively, and with due 

 regard to the nature of the evidence upon which they are based. 



10. The Preliminary Experiments on ss. Ascanius. 



Various changes in the disposition of the aneroid were made dui-ing 

 the voyage, and additions were introduced as experience was gained ; for 

 example : (1) the instrument was mounted on the support designed for it 

 instead of being allowed to rest on the table, an advantage clearly shown 

 in fig. 12 ; (2) oil damping was substituted for air damping ; (3) a level 

 and sliding weights were added to enable the instrument to be adjusted 

 horizontally whenever necessary. Discontinuities were thus introduced 

 which probably account for the discrepancies in the harbour station 

 observations in Cape Town, Fremantle, and Adelaide (fig. 3). On account 

 of these the reduced results are scarcely of sufficient value to justify a 

 description of them in further detail than is conveyed in figs. 12 

 and 13, the data for which have been obtained from the first method of 

 reduction described above : — 



(1) The low values of the north latitude observations, fig. 12, are due to 



