September 5, 1890.] 



SCIENCE. 



137 



same effect. Inasmuch as these two foreign discussions reached 

 conclusions directly contradictory to all the teachings of meteor- 

 ology, the importance of determining where the error lies, and of 

 establishing the truth, will readily be seen. 



A careful study of the question will show that the entire ditB- 

 culty, and apparent contradiction, has arisen from a neglect of the 

 consideration that at considerable heights in the atmosphere a 

 lower temperature has a tendency to contract the air and cause a 

 diminution of pressure, and a higher temperature just the rever.=:e. 

 The best example of this is to be found at Pike's Peak (14,134 

 feet), where the lowest pressure ever recorded was 16.88 inches, on 

 Jan. 20, 1883, while the temperature at the summit was— 34°, and 

 while a high area of great magnitude was passing at the base. 

 This shows that we must ignore fluctuations in pressure at the 

 high station, and consider only those below. When this is done, 

 the whole difficulty vanishes at most stations. Dr. Hann seems 

 to have found a few cases where an increase of pressure at the 

 lower stations near Sonnblick has been coincident with an in- 

 crease of temperature on the mountain 8,700 feet above. It may 

 be well to pursue this discussion under a slightly different form, 

 £md unite the results as obtained at Mount Washington (6,279 feet) 

 with those in Austria. 



The plan proposed is simply to compare side by side the tern 

 peratui-e fluctuations at both base and summit. If we had balloon 

 observations at 10,000 feet and others at sea-level on the earth's 

 surface, such conipai-isons could be made readily and accurately; 

 but it should be noted that when we use mountain observations, 

 especially those on ranges and not isolated peaks, we cannot hope 

 for an absolute comparison. The difficulty will be enhanced if 

 our base station lies at some distance from the mountain, though 

 in the case of an extended range we may obviate some of this 

 source of error by taking stations on both sides of the range. It 

 will be universally admitted that, north of the equator, the usual 

 fluctuations of temperature at sea-level on the passage of storms 

 and high areas are perfectly well known, though these may at 

 times be masked or even reversed, as, for example, when the cen- 

 tre of the storm passes just a little south of the station. In gen- 

 eral, as a storm comes up, there is a southerly breeze and a great 

 increase of temperature. This increase of temperature is observed 

 even though there be a calm and the sky be clouded, hiding the 

 sun's direct rays. We naturally conclude that this heat condition 

 is an accompaniment of the storm, and is largely independent of 

 the sun's direct influence in raising the temperature. Exactly the 

 reverse of this is experienced when a high area or a clearing con- 

 dition approaches a station. Here the sky is perfectly clear, and 

 though the sun has apparently a much better opportunity to heat 

 up the earth and air, still we find a marked lowering of tempera- 

 ture. This is the normal condition, but suppose we find that with 

 the increased pressure there is increased temperature at sea-level, 

 or that the clouds come up and there is rain, then we must con- 

 clude that the conditions are abnormal, and in any general discus- 

 sion or comparison of temperature conditions at the base and 

 summit of a mountain we must give such eases a separate study 

 and not unite them with the normal fluctuations. It seems quite 

 plain that we may draw curves showing the observed tempera- 

 tures at base and summit, and compare them directly. There is 

 a slight difficulty, however, which must first be overcome, and it 

 is this. At the earth's surface there is a marked daily effect from 

 the sun's direct heat which generally causes a steady rise of tem- 

 perature from sunrise to about 3 p m., and this would mask the 

 other conditions. It would be a great advantage if we could use 

 observations more than once a day, as the maximum point in the 

 passage of a storm and a minimum point in a high area might oc- 

 cur at any hour of the twenty-four. One way of eliminating this 

 diurnal range would be to apply the difference between the 

 monthly mean and the mean for any hour to each daily observa- 

 tion of that hour; for example, we would have to add a little 

 to nearly every sunrise observation and subtract from nearly 

 every maximum observed; but a better way still would be 

 to take the mean of the hour which agrees most closely with the 

 mean for the day and apply the difference between that and the 

 hourly mean to each observation, as this would save one-third of 

 the labor when we are studying three observations each day. In 



the latter case the 9 p.m. observation, for example, would be 

 projected without modification. In projecting the temperature 

 curves it was found most convenient to use the night observation 

 rigidly and to interpolate the morning and afternoon observations 

 if either or both differed widely from that. For Mount Wash- 

 ington the station at Burlington was chosen for the base until it 

 was closed in 1883, and after that Portland. After projecting the 

 curves there was found to be a most extraordinary similarity be- 

 tween the changes at the base and summit. To illustrate this I 

 have drawn Fig. 1, which gives the fluctuations for January, 1876. 

 I think this will be recognized as a perfect accordance. The slight 

 hitch on the 16th on the summit has only a very slight bending at 

 the base, but it is noticeable. Such slight coincidences were ig- 

 nored in the summing-up. In the 78 colder months there were 

 1,128 accordances out of 1,240 cases, or 91 per cent; and in the 

 warmer months, April, 1873, to September, 1879, 42 months, there 

 were 553 accordances in 601 cases, or 93 per cent. It would seem 

 that even if there were no explanation for these few discordances, 

 the evidence is conclusive that; whatever fluctuations of tempera- 

 ture take place at the base, they are faithfully repeated at the 

 summit. The comparison will seem all the more effective when 

 we reflect that this thin strip between these curves represents a 

 difference in vertical height of over 6,000 feet. 



I have made a careful study of the discoi-dances, and find that 

 they can all be explained under the following heads: (1) Often the 

 curve turns at the summit before it does at the base, or, in other 

 words, a lagging at the base causes the summit curve to cross 

 that at the base. (3) At other times there is an abnormal condi- 

 tion of the upper atmosphere, a fall of rain, for example, in the 

 centre of a high area, which shows a remarkable disturbance of 

 temperature conditions. (3) There is sometimes in a high area 

 a perfectly clear sky, which promotes intense radiation from the 

 soil at the base, but which has no counterpart at the summit. 

 For example, during the progress of a high area which culminated 

 at 4 P.M., Nov. 16, 1874, the following temperatures were ob- 

 served ; — 



Here the minimum temperature at the summit was reached 

 during the afternoon of the 16th, but the great radiation after 

 dark at the base gave a minimum at that point at night, though 

 it is also probable that a portion of this was due to a lagging at 

 the base. (4) There are occasions in the centre of a high area, 

 which has only a very slight onward motion, when the sun's 

 heat appears to have an abnormal effect upon the ah* col- 

 umn, causing an increase in temperature at the summit above 

 that at the base. In all the cases examined there were only tvna 

 under this last head. There is no difficulty at all in explain- 

 ing all the exceptions, and these may fairly be said to prove 

 the rule. The evidence is overwhelming that if the principles 

 just laid down are accepted there is a marked increase in tem- 

 perature at 6,000 feet height in our storms, and a decrease in our 

 high areas. 



It is a very interesting fact that the distance between the curves 

 at the base and summit of Mount Washington during the passage 

 of storms and high areas seems to be nearly constant. The slight 

 crowding on the right-hand side of the curves between Nov. 6 and 

 11, and 26 and 30, is due to the fact that the base is to the north- 

 westward of the summit, and hence the latter lags behind a little 

 as a storm moves to the east. Does not this similarity prove that 

 there is no uprush of air in the storm, nor a downrush in the high 

 area? 



To complete this investigation it is necessary to make a similar 

 study of the observations used by Dr. Hann in arriving at his con- 

 clusions. It should be borne in mind that Sonnblick, the moun- 

 tain station used by Dr. Hann, is almost entirely outside of the 



