JrxE 25, 1886.] 



SCIEXCE. 



571 



wider acceptance than our knowledge of the facts 

 would seem to justify. 



Assuming the continents to be simply so much 

 matter, of half the earth's mean density, on the sur- 

 face of our otherwise closely centrobaric spheroid, 

 it may be shown that individually they will draw 

 the sea-surface up towards their centres by consider- 

 able amounts (about a thousand metres at most), 

 leaving corresponding though not equal depressions 

 opposite those centres ; and that collectively they 

 will produce a wavy sea-surface, in which the max- 

 imum radial distance from crest to hollow is about 

 six hundred metres. The theory, and the equations 

 assigning the form and position of this wavy 

 surface, have been developed by Helmert in his ex- 

 cellent treatise on geodesy (' Die mathematischen und 

 physikaliscben theorieen der hoheren geodasie '), 

 from which the above figures have been taken. If 

 we dropped our examination of the question at this 

 point, we might infer the reality of the wavy surface 

 just described. The existence of such assumed con- 

 tinents, however, implies a proportionate variation 

 of gravity along the sea surface and along the same 

 level surface extended through the continents. They 

 would, for the most part, produce an excess of 

 gravity over the continental and a deficiency over 

 the sea areas. But this conclusion is in direct con- 

 tradiction with the results of pendulum experiments. 

 The assumption, therefore, that the continents are 

 superficial masses, unbalanced in their attractive 

 effects, is, as clearly shown by Helmert, inadequate, 

 and must, together with the conclusions based there- 

 on, be modified or rejected. 



Some writers, notably Fischer and Listing, have 

 proved the existence of a highly irresrular sea-sur- 

 face by a still more unsound process than that in- 

 dicated above would be if we neglected to examine 

 its fundamental assumption. This process, in brief, 

 rejects in an equation a term of the same order as 

 those retained, and arrives at a simple relation be- 

 tween the variation of gravitv and the radial dis- 

 tance from the actual sea-surface (or geoid) to the 

 mean spheroidal surface. Helmert fitly character- 

 izes this relation as entirely worthless (ganz wertlos), 

 since it fails in every case to give the proper sign 

 when the increments of gravity and radial distance 

 due to the combined action of the continents are 

 substituted in it. 



Those desiring to examine minutely the merits of 

 this question should consult the above-named treatise 

 of Helmert, who gives a critical review of the 

 cognate works of Fischer, Listing, Bruns, and 

 others. For the benefit of the general reader, it 

 may be stated, that, although the sea-surface is un- 

 doubtedly somewhat irregular, geodesy and geology 

 have as yet furnished no adequate evidence of 

 irregularities amounting to more than ten metres. 

 Additional information, of which it must be ad- 

 mitted there is great need, may disclose the existence 

 of a surface having hills and hollows separated by 

 an interval of fifty or possibly a hundred metres ; 

 but irregularities of any greater extent appear to be 

 quite improbable. 



The suggestion of your correspondent, that the 

 barometer would indicate any large elevations or 

 depressions in the sea-surface, is not well grounded. 

 The surfaces of equal pressure in the atmosphere 

 must approximate to parallelism with the sea-sur- 

 face, however irregular it may be. In a state of 

 quiescence the air-surface in contact with the sea 



is necessarily a surface of equal pressure. The 

 barometer would therefore, if moved from one point 

 to another along the sea-surface, register only such 

 variations in pressure as are due to changes of tem- 

 perature, winds, etc., and hence afford no indication 

 of the elevations and depressions in question, if they 

 exist. E. S. Woodward. 



Washington, DC, June 17. 



3arometer exposure. 



Mr. Clayton's letter concerning the influence of 

 wind on the indication of the barometer broaches a 

 subject of great importance to theoretic and practical 

 meteorology, and I trust it may lead to the execution 

 of the experiments essential to the intelligent treat- 

 ment of the difficulty. As his conclusions are called 

 in question by President LeConte, I take the liberty 

 of rehearsing some investigations of my own which 

 tend to sustain Mr. Clayton's conclusions. 



In June, 1873, an elaborate series of synchronous 

 barometric observations were made by the signal 

 office at four stations on the summit and slope of 

 Mount Washington. In testing a special method of 

 barometric hypsometry , I had occasion to discuss these 

 observations, and I discovered an important anomaly 

 which was correlated with the velocity and direc- 

 tion of the wind. The discussion cannot be repeated 

 here, for lack of space ; but it may be said that its 

 method and material were such as to leave no rea- 

 sonable doubt that the wind was the disturbing fac- 

 tor, while they afforded quantitative results far more 

 precise than can be reached by any method of reduc- 

 tion to sea-level. The reader who cares to examine 

 them should consult the 1 Second annual report of 

 the U. S. geological survey,' pp. 521-534 and 562-565. 

 One of the specific conclusions was, that a north- 

 west wind of fifty miles per hour, by drawing air 

 out of the summit observatory, presumably through 

 the chimney, caused the mercury in the barometer 

 to stand .13 of an inch too low ; and it was estimated 

 that a wind-velocity of a hundred miles would lower 

 the mercury more than half an inch. 



I think President LeConte is mistaken in suppos- 

 ing that the matter could be simply tested by com- 

 paring the indications of a barometer in a room with 

 those of a barometer out of doors. If the out-of- 

 door barometer were placed on the windward side of 

 a building or other obstruction, and close to it, it 

 would be immersed in compressed air. and read too 

 high. If placed under the lee of an obstruction,, it 

 would be surrounded by relatively rarified air, and 

 read too low. If placed in a position uninfluenced 

 by obstructions, the locus of difficulty would be trans- 

 ferred from the surrounding atmosphere to the 

 instrument itself, for the air chamber above the 

 mercury in the cistern of the barometer would itself 

 be influenced by the wind so as to receive a tension 

 abnormally high or low. These statements, based 

 on familiar physical laws, are not individually sus- 

 ceptible of ready verification, because,, while the wind 

 blows, all local tensions are disturbed, and we have 

 no standard air-pressure for comparison. I have, 

 however, determined experimentally that the read- 

 ing is higher in front of an obstruction than behind 

 it. A difference of .15 of an inch was found be- 

 tween barometer-readings on opposite sides of the 

 apex of an acute mountain-peak. 



In my opinion, the proper method of escaping the 

 difficulty is, not to place the barometer out of doors, 

 where observation during a wind is itself a matter 



