SCIENCK 



[Vol. XV. No. 36S 



of considerable importance. The wedges should press out the 

 mouth of the clamp slightly, but more and more towai'ds the 

 back. Direct application of the loads proved quite unsuited, as 

 zinc is highly influenced by the rapidity of the changes. Pro- 

 fessor Martens, therefore, resorted to a testing-machine of his 

 own design, three different modifications of which were em- 

 ployed. As indicators for these apparatus, a circular- vessel 

 filled with mercury was employed, from the side of which a 

 vertical tube branched off. The cover of this vessel wa» formed 

 by a sti'ong centi-al plate supporting a weight surrounaea by a 

 ring of german-silver. The strain imparted to the test-piece was 

 partly taken up by the weight, the mercury column effecting the 

 balance. This arrangement, which resembles others employed 

 for similar purposes, did not answer ; it was, moreover, not self- 

 recording. The mercury -tube was therefore replaced by a hori- 

 zontal cylinder with a piston- rod ended in another piston moving 

 in a second cylinder with a slide-valve, which was actuated by 

 an electric device comprising electro-magnets and relays. The 

 common piston-rod carried a pointer recording on a paper diaim. 

 A third device, also electrical, but worked by gTavity instead of 

 ■water-pressure, was employed for the highest loads up to 50,000 

 kilograms. These three arrangements labored under the disad- 

 vantage that the cover of the mercury vessel retained an amount 

 of mobility suificient to affect the accuracy of exact measure- 

 ments. Professor Martens hence returned to an often-employed 

 arrangement, utilizing the elasticity of a spring of an elastic 

 steel rod. The idea is, that the variations of the rod are marked 

 directly (and without being magnified by multiplying levers or 

 other devices whose acc.uracy Mr. Martens altogether questions) 

 by means of a little conical diamond point on glass plates of the 

 size for microscopic slides, fixed on a platform moved by means 

 of a micrometer-screw and adjusting-spring, in a direction at 

 right angles to that of the axis of the rod. Two of the resulting 

 curves would occupy a space of not more than a sqviare milli- 

 metre. The plates were examined and measured in a large Zeiss 

 microscope provided with micrometers for both object and ocular 

 glasses In this form, the recording device has been consti-ucted 

 by Mr. Boehme. It is, however, intended to leave the platform 

 at rest, and to register the movements in the direction of both 

 the abscissae and the ordinate. 



The chief objects of the tests were to ascertain the elasticities 

 at ordinary temperatures and at SO", 120'', 150'^, 170°, and 200° 0. 

 (between 176° and 392° F.), and to ascertain the influences of dif- 

 ferent modes of rolling and of time-effects The latter are strik- 

 ing. One can hardly speak of the elasticity of rolled zinc, as- 

 even under very small strains the permanent expansion varies 

 with each change of load. There, was always a noticeable after- 

 stretching. Wlien cold, the breaking strength was 33 per cent 

 larger, the breaking extension 22 per cent smaller, and the ' 'ful- 

 ness degree" (i.e., the ratio of the area comprised by the curve 

 to the rectangle formed by the greatest extension multiplied by the 

 greatest force) neither larger nor smaller, in a direction at right 

 angles to the rolling, than in that of the rolling. The two sam- 

 ples supplied by other works showed, however, opposite charac- 

 teristics, and one test-piece particularly deviated in a manner 

 probably to be accounted for by some peculiar ti'eatment during 

 manufacture, the chemical composition seeming to afford no ex- 

 planation. Rising temperatures modifled the results. The break- 

 ing strength increased considerably in thinner sheets ; that is, in 

 such as have undergone greater and more continued pressure in 

 the rolls. It rose from 11 kilograms per square millimetre for 6- 

 milUmetre plates, to 19 kilograms for plates .48 of a millimetre 

 thick. The English equivalents of these values are 17.5 and 30 

 tons per square inch respectively for plates of .24 and .019 of an 

 inch in thickness. The breaking extension decreases fh-st, and in- 

 creases rapidly afterwards. For the temperature tests, the pieces 

 were heated in a linseed-oil bath. The results confu-m the well- 

 known and important fact, fii-st established by Silvester and Hob- 

 son of Sheffield, that zinc should be worked, rolled, stamped, 

 turned, etc. , at 800° F. , and that any higher temperature should 

 carefully be avoided. On the whole, the tests demonsti-ate clearly 

 that ordinarily tensile sti-ength tests are not alone sufficient, and 

 should be combined with folding and bending tests. 



METEOROLOGICAL OBSERVATIONS ON PIKE'S PEAK. 



Since the Boy den fund of the Harvard College Observatory 

 was established for the purpose of obtaining astronomical ob- 

 servations at some station of great elevation above the level of 

 the sea, an inquiry into the meteorological character of such 

 stations seemed desirable before undertaking the proposed work. 

 It was known that a long series of meteorological observations 

 at the highest station ever permanently occupied for such a 

 purpose had been made by the United States Signal Service on 

 the summit of Pike's Peak, in Colorado. It was accordingly 

 proposed to the chief signal officer of the United States Army, 

 Gen. A. W. Greely, that these observations should be printed 

 at the expense of the Boyden fund, in the "Annals of Harvard 

 College Observatory ; ' ' and his courteous co-operation has en- 

 abled this plan to be carried out, as shown in Vol. XXII. of the 

 "Annals" of the observatory, just published. 



The summit of Pike's Peak, Colorado, is situated in latitude 

 38° 50' north, longitude 105° 2'west, and has a height of 14,134 

 feet above sea-level, as determined by spirit-level from Col- 

 orado Springs. It is the highest meteorological station in the 

 world; Leh, Ladakh, being 11,503 feet, and the Sonnblick, 

 Austria, 10,154 feet. The station on the summit of Pike's 

 Peak was established in October, 1873, and the first telegraphic 

 report sent on Nov. 6 of that year. The telegraph line was 

 frequently interrupted, and for long periods, until November, 

 1882, when it was virtually abandoned, owing to the great cost 

 and the difliculty of its maintenance. Observations, however, 

 were continued until September, 1888. 



During the first few weeks the observations were more or less 

 interrupted, and it has been deemed best to commence the pub- 

 lication from Jan. 1, 1874, at which date the station was in 

 complete working order. 



Pike's Peak rises very abruptly from the eastward, being about 

 8,000 feet above Colorado Springs, which is within ten miles or 

 so from the summit. The open plain extending to the east- 

 ward affords unusual advantages for noting such cloud and 

 storm phenomena as originate or move to the eastward of the 

 mountain ; and even the peaks to the westward are enough 

 lower to permit observation of storm and cloud conditions 

 below the level of the observer on Pike's Peak. 



Perhaps the most notable fact resulting from a cursory ex- 

 amination of the meteorological elements is the remarkable 

 resemblance between the recurring annual phases of atmospheric 

 pressure and the temxjerature of the air. The curves of these 

 elements not only are alike in having a single bend, but the 

 maximum phase of both occurs in July, and the minimum in 

 January. Not only are these elements coincident in their 

 extreme phases, but the annual march is the same ; so that the 

 two curves are not only parallel, but almost coincident. When 

 examined mathematically, it will be seen that not only are the 

 plus and minus changes from month to month the same for 

 both elements, but they bear a close, definite, and apparently 

 dependent relation to each other, tbe mean monthly pressure 

 rising or falling about .016 of an inch for each change of one 

 degree Fahrenheit in the monthly mean temperature. 



A similar relation between the mean monthly pressui-e and 

 mean temperature obtains on the summit of Mount Washing- 

 ton, New Hampshire (elevation 6,279 feet above the level of 

 the sea) ; but the barometer and temperature curves for the 

 year at this last-named station are not as regular as on the 

 summit of Pike's Peak. On Mount Washington, while the 

 extremes of monthly mean temperature fall likewise in Janu- 

 ary and July, yet tbe maximum monthly pressure shows a 

 tendency to prolong itself into August, and the minimum 

 pressure to continue throughout January, February, and March. 

 The relation on Mount Washington of monthly changes of 

 pressure to like changes of mean temperature differs slightly 

 fi-om that of Pike's Peak, being about .012 of an inch rise or 

 fall for each degree Fahrenheit. 



The actual atmospheric pressure at Rocky Mountain stations 

 above 4,000 feet elevation attains its minimum in January and 

 its maximum in July or August ; and the barometric phases 



