518 



SCIENCE. 



[N. S. Vol. III. No. 66. 



thing which he has at his finger's end to make 

 the distinction between the necessary and the 

 sufficient condition for the truth of a statement, 

 and there is no reason why other scientists 

 should not speak with the same precision. One 

 thing is the necessary condition for the truth of 

 another, if the latter cannot be true in its ab- 

 sence ; it is the sufficient condition, if it must 

 be true in its presence. It may be matter of 

 question whether ' test of truth ' should be 

 used in the sense of necessary or of sufficient 

 condition of truth, but it certainly should not 

 be used in both senses in the same sentence. 

 ' Evidence ' is the sufficient condition for the 

 truth of a statement, but it is not in every in- 

 stance necessary. I need no evidence to con- 

 vince me that I am conscious. Now those who 

 regard conceivability in the way that Prof. 

 Brooks objects to, do not for a moment consider 

 it to be a sufficient condition of the truth of 

 any statement, but they do consider it to be the 

 necessary condition of the truth of every state- 

 ment. The inconceivability of a statement is for 

 them the sufficient test of its falsity, and its 

 conceivability is the necessary test of its truth. 

 Instead of saying, therefore, with Prof. Brooks, 

 that the test of truth is evidence and not conceiva- 

 bility (a statement which gives me a slight feel- 

 ing of dizziness), it would be better to say that the 

 test of truth is evidence, and inconceivability is no 

 criterion (or test) of falsity, provided the exact 

 terms, necessary and sufficient, should be con- 

 sidered too pedantic. 



I have used the terms necessary and sufficient 

 because they have been consecrated to this 

 purpose by the mathematician, but I believe 

 that essential and sufficient, or perhaps requisite 

 and sufficient, would convey the meaning much 

 better for ordinary language. We shovild then 

 say, evidence is a sufficient test * and conceivability 

 ■is not a requisite test of truth. The sentence 

 " conceivability is not a necessary test of truth " 

 is somewhat ambiguous; it might mean ' is not 

 a test such that the truth necessarily follows 

 from it,' instead of 'is not a test which it is 

 necessary to have fulfilled if the truth is to 

 hold.' But ' requisite test of truth ' is not open 

 to any ambiguity. 



* That, for nearly all truths, evidence is also a re- 

 quisite test, is true, but is denied by no one. 



I am convinced that if the terms requisite and 

 sufiicient (or something equivalent to them) 

 were to come into common use as defining the 

 kind of ground, reason, argument, condition or 

 test that the writer has in view, it would con- 

 duce very much to facility of comprehension on 

 the part of the reader. M. M. 



THE TEMPEEATUEE OF THE EAETH'S CEUST. 



Me. Seeeno E. Bishop, in his letter in Sci- 

 ence, March 13th, remarks that it would be 

 interesting to ascertain what are the rates of in- 

 crease of temperature now under regions where 

 the subsoil is permanently frozen, as in the 

 tundras of Siberia and Alaska. 



Attention may here be called to the Report 

 made to the British Association in 1886, by the 

 committee appointed to organize a systematic 

 investigation of the depth of the permanently 

 frozen soil in the polar regions. Of some 

 twenty-two localities mentioned in that Report, 

 Jakutsk, Siberia, lat. 62°, is perhaps the most 

 noteworthy, the limit of the frozen soil being 

 620 feet and the temperature rate 1 ° for 28 feet. 



The transcendental formula employed by 

 Lord Kelvin in his well-known chapter on the 

 ' Cooling of the Earth ' furnishes results ia 

 marked harmony with the temperature rate as 

 determined by many observations. (Prestwich, 

 Proceedings of the Royal Society, 1886.) It 

 does not logically follow, of course, that Lord 

 Kelvin ' s premises are necessarily correct. How- 

 ever, whether we accept the argument in the 

 ' Cooling of the Earth ' or rely on observations 

 alone, we must for the present regard 1° F. per 

 50 feet (approximately) as expressing the law 

 of the rate of increase of the temperature of the 

 earth's crust near the surface ; some local factor 

 should be looked for as the cause of such an ex- 

 ceptionally low rate of increase as that found in 

 the Calumet mine, or such a high rate as that 

 in the Jakutsk mine. In any case it is scarcely 

 safe to assume, as Professor Agassiz seems to 

 do, that the rate observed to the bottom of the 

 Calumet mine holds to the depth of 19 miles 

 and beyond, and thence to conclude that the 

 earth's crust has a thickness of 80 miles. The 

 crust of the Lake Superior region may ha\'e 

 counterbalancing abnormal features, so that the 

 low temperature rate for the first mile is amply 



