QC J 3 rp UC-NRLF W48 B M 253 533 George Davidson i <~n i HOW TO USE THE ANEBOID BABOMETEB BY EDWARD / WHYMPEE I. COMPARISONS IN THE FIELD II. EXPERIMENTS IN THE WORKSHOP III. UPON THE USE OF THE ANEROID BAROMETER IN DETERMINATION OF ALTITUDES IV. RECAPITULATION NEW YORK CHARLES SCRIBNER'S SONS, 743, 745 BROADWAY 1891 All rights are reserved GIFT ZOOLOGICAL COLLECTIONS FROM THE GREAT ANDES OF THE EQUATOR. WALTER BURTON, Naturalist, of 191 WARDOUR ST., LONDON, W., begs to inform Collectors of BIRDS, INSECTS, REPTILES, and FISHES that he has the collections formed l>y EDWARD WHYMPER, Es^., upon his Journey in Ecuador for sale, including : The first set of the DIURNAL LEPIDOPTERA, classified in 14 boxes, embrac- ing 240 specimens of 106 Species, taken at various elevations up to 16,000 feet above the sea, with Localities and Descriptions attached ; duplicates of Colias dimera, Doubl. and Hew., Picris xanthodue, Lucas, Pieris elodia, Boisd., etc., Is. 6d. to 2s. each ; numerous skins of Humming and other BIRDS, in the very best condition ; excellent specimens of the very rare and interesting FISH, Cyclopium cydopum, Humboldt (see Travels amongst the Great Andes, chap, xii, and the description in Supp. App. by the late Dr. F. Day), 4s. to 7s. each ; amongst the FROGS perfect examples of males, females and young of the curious Nototrema marsupiatum (Dum. & Bibr.) ; also of the new species Hylodes IVTiymperi, described by Mr. G. A. Boulenger, captured from 11,000 to 13,200 feet; Hylodes conspicillatus, Gtlir. ; H. unistriyatus, Gthr., and young; Phryniscus Icevis, Gthr. ; Dendrohates tinctorius (Schneid.) ; Bufo ayua, etc. etc., 3s. to 7s. 6d. each. Fine specimens of the LIZARDS Liocephalus trachycephalus, L. iridcscens, Gthr. ; Ecpleopus (Pholidobolus) montium, Peters ; etc. etc., 5s. to 7s. each. Examples of the SNAKES Liophi's regime, L., var. albiventris, Jan. ; L. splendent, Jan. ; Bothrops atrox, (L.) ; R Schleyeli, (Berthold) ; Leptoynathus nebulatus, (L.) ; Elaps Marcyrari, (Wied.), var. ancolaris, Jan. ; Oxyrhopus clcelia, Jan.; Boa constrictor, L., etc. etc., 5s. to 7s. each. Specimens of the SCORPIONS Broteas subnitens, Gervais (from 10,000 feet and upwards), and Centrums Americaims, De Geer ; of the AMPHIPOD Hyalella inermis, S. I. Smith (obtained from 12,000 to 13,000 feet); and the fine BUG Carineta basalis, Walker. Good series of the new species of ANT Phcidole monticola (females, soldiers and workers) described by Mr. P. Cameron; and specimens of Camponotus atriceps, Sm., G. sylvaticus, Oliv., Atta sexdens, L., Pseudomyrma gracilis, Fabr., Edatomma quadridens, Fab., and Pachycondyla carbonaria, Smith, etc. etc., Is. to Is. 6d. each. And a numerous series of the special COLEOPTERA of the highest zones of the Ecuadorian Andes, including the following New Species which are described in the Supplementary Appendix to Travels amongst the Great Andes of the Equator: Leucopeltea albescens, Bates ; Colpodes stcno, Bates ; C. alticola, Bates ; Pterostichns integer, Bates ; Aniso- tarsui bradytoides, Bates ; Pelmatellus Andium, Bates ; P. variipes, Bates ; P. oxynodes, Bates ; Astylus bis-sexyuttatus, Gorham ; Ananca dcbilis, Sliarji : Meloe sexguttatus, Sharp; Sterculia impressipennis, Sharp; Epitragus (?//'//'.<, Sharp ; Philonthus Whymperi, Sharp ; Silpha microps, Sharp, etc. etc., Is. 6d. to 2s. 6d. each. , WALTER BURTON, 191 WARDOUR STREET, LONDON. HOW TO USE THE ANEROID BAKOMETER BY EDWARD WHYMPER H I. COMPARISONS IN THE FIELD II. EXPERIMENTS IN THE WORKSHOP III. UPON THE USE OF THE ANEROID BAROMETER IN DETERMINATION OF ALTITUDES IV. RECAPITULATION NEW YOEK CHAELES SCEIBNEK'S SONS, 743, 745 BEOADWAY 1891 All rights are reserved I A qe W PBEFACE. THE following paper is divided into four sections. The first records com- parisons of the aneroid against the mercurial barometer in the field ; the second is concerned with experimental research in the workshop ; the third is occupied by practical considerations arising from the facts recorded ; and the fourth is composed of a recapitulation of the principal points which are dwelt upon in the previous parts, and forms an index to the whole. The investigations which are recorded in these pages have extended over eleven years. In 187980, having occasion to pass some length of time at great elevations in the Equatorial regions of South America, I took the opportunity to make comparisons of the aneroid against the mercurial barometer at low pressures. Some of these comparisons are given in the First Part. 1 On entering upon this journey, I entertained the hope that close approxi- mations to true atmospheric pressure might be obtained by employing several or a number of aneroids. The idea was that the plus errors of some instru- ments might balance the minus errors of others, and that means of the whole might come near the truth. This hope was speedily dissipated. Each individual instrument indicated lower pressures than the mercurial barometer, and means of the whole were, consequently, less than the truth. Deduced altitudes were much too high. In strong contrast to these un- welcome facts there were others equally perplexing, although less unpleasant. It was found that even when the aneroids had acquired minus errors of one to two inches they might on differences of level of several thousand feet indicate nearly the same differences of pressure as the mercurial barometer. Ascending 1 Part 1 is reprinted from the Appendix to Travels amongst the Great Andes of the Equator. This volume contains an account of the journey, and describes the circum- stances under which the comparisons were made. These two works are published simul- taneously, and can be obtained separately. B W M294554 2 PEE FACE. observations of aneroids, it was found, never agreed with descending ones, and the latter always indicated less difference of pressure and less difference of level than the former, if they were made only a short time after them ; but, in such cases, means of the ascending and descending observations closely accorded with means of similar observations made with the mercurial barometer. A long time elapsed after my return from this journey before I commenced to find a solution of these puzzles. I examined and rearranged my observa- tions for several years before it occurred to me to tabulate the means of the whole of the aneroid readings in chronological order, to do the same for the readings of each individual instrument, and to take out the differences from the mercurial barometer of each mean and of each individual reading. When this was done I saw that the results were influenced by the lapse of time between the readings as well as by variations in pressure. The general con- clusions at which I then arrived are stated upon pp. 9, 14. Having introduced the matter to the notice of Mr. J. J. Hicks, of 8 Hatton Garden, and gained his co-operation, I then proceeded to test aneroids under varied conditions of time and pressure ; in the end finding explana- tions for the points of perplexity, and that aneroids, whilst largely differing amongst each other, all follow rules which have the force of law. The prosecution of these experiments has extended over a far greater length of time than appeared necessary at their inception, and has absorbed almost all my leisure during the last few years. The labour will have been well bestowed, and I shall attain my aim, if in these pages I succeed in explaining some points in the behaviour of aneroids which have undoubtedly been puzzling, and have seemed contradictory and anomalous ; and thus render it possible for the ever-increasing number of those who employ these instruments in the field to use them with more confidence and certainty. EDWARD WHYMPER. PART I. COMPARISONS OF THE ANEROID AGAINST THE MERCURIAL BAROMETER. Aneroids were carried to Ecuador to endeavour to ascertain whether the 'means of the readings of several, or of a number, would or would not accord with the mercurial barometer at low pressures. It has long been known that the indications afforded by a single aneroid are apt to be of a very deceiving nature, even at moderate elevations ; but it seemed to me possible if several, or if a number were employed, that one might, by inter-comparison, discriminate between those which went most astray and those which held closely together ; and that, by adoption of the means of the readings of the latter, a decent approximation might be obtained to the truth, possibly even at great altitudes. It may be added that I wished this might prove to be the case ; for the portability of aneroids, the facility with which they can be read, and the quickness of their action, would render them valuable for many purposes, if their indications could be relied upon. Eight aneroids were taken. One of these, by Casella, marked No. 580, had been made for an earlier journey ; and, through being only graduated to 20 inches, was of no service for comparison at great heights. 1 The seven remaining aneroids were constructed expressly for the expedition, and were under trial and examination for nearly twelve months before our departure. They were selected from picked instruments, and only those were taken such as were, so far as one could tell, in all respects as perfect as could be produced. 2 These seven aneroids were marked A G- A, B, and C were graduated from 31 down to 15 inches, and D, E, F, G were graduated from 31 down to 13 inches, a range which I thought would be sufficient for my purposes. 3 It became apparent at an early stage of the journey, a. that the whole of the aneroids had acquired considerable errors ; &. that they differed amongst each other to a very large extent ; and c. that neither means of the whole, nor means of those which held closest together, nor means of any combination, would give decent approximations to the truth. The more evident this became the greater importance I attached to the preservation of the mer- curials. Comparisons of the aneroids against the mercurials were neverthe- 1 This was left with Mr. Chambers at Guayaquil, as a reserve for him, in case accident befell the Standard Mercurial ; and he read both instruments during the whole of our absence in the interior. 2 I abstain, however, from mentioning the names of the makers (to whom I am much indebted), lest the remarks which follow should be construed to their disadvantage. 3 It proved to be inadequate. 4 COMPARISONS IN THE FIELD. less continued until the end of the journey - 1 and after two months' experience in the interior the behaviour of the aneroids in ascending and descending was so well ascertained that one might, I think, have made use of their indications to determine differences of level without committing very great mistakes. . In the following pages, I propose first to give some of my experiences, and then to draw such conclusions as appear to be warranted. 2 1. Shortly before my departure from London, I made (on October 25, 1879) a final comparison of the aneroids against the mercurial barometer. Only one of the aneroids corresponded exactly, and of the others some were too high and some were too low, the greatest difference between them amounting to 0*225 of an inch, and the mean of the whole showing an error of +0-148 of an inch. 2. Our ship stopped a clear day (November 20, 1879) at Jamaica, and I took the opportunity to carry the aneroids to the top of the Blue Moun- tains, 3 comparing them against the mercurial before starting and upon return, and comparing them against each other at the highest point attained. The following is the complete record, and it will be seen from it that the mean No. of Barometer. At start. At top. On return. Aneroid 580 29-980 25-430 29-975 A ... 29-850 25-500 29-850 B . . . 29-800 25-250 29-800 C . . . 29-700 25-120 29-650 D . . . 29-850 25-400 29-880 E . . . 29-800 25-300 29-750 F . . . 29-750 25-350 29-700 G . . . 29-800 25-310 29-700 Extreme differences 0-280 0-380 0-325 Mean of aneroids 29-816 25-332 29-788 Merc. bar. No. 554 . 29-876 . 29-854 Mean error of aneroids -0-060 -0-066 error of the aneroids (which in London was + 0-148) had changed on arrival at Kingston to - 0*060, and upon return in the evening of November 20 it was still further increased to 0-066. 1 In all the comparisons which are made throughout this paper the readings of the mercurial barometer are reduced to 32 Faht. 2 Paragraphs 1-10 should be read in connection with the tables at pp. 10, 11. 3 Drove to Gordonstown, walked thence to Newcastle, and on until we came to a notch in the mountains commanding a view of the northern side of the island. Read the aneroids at this place. COMPARISONS IN THE FIELD. 3. Comparisons were again made at Colon and Guayaquil, 1 and at neither of these places (at the level of the sea) was there any further in- crease in the mean error of the aneroids ; but at Munapamba (where we commenced to mount the outer Andean ranges) their difference amongst each other had risen to 0*500 of an inch, and the mean error was increased to -0-098 of an inch. 4. At Tambo Gobierno (the culminating point of the road over the outer Andean range), 10,417 feet above the sea, the mean error of the aneroids had risen to 0*487 of an inch, and the extreme difference of their readings amounted to 0*715. 5. With the descent on the other side the aneroids came more closely together, but their mean error continued to augment, being upon arrival at Guaranda 2 (8894 feet) -0*520, and it rose in one week to 0*655. The ' greatest difference ' also continued to increase, and it stood on Christmas Day at 0*800 of an inch. The following record, showing the continual increases in the errors, will be found interesting. No. of Barometer. Dec. 18, 1879. Dec. 20, 1879. Dec. 23, 1879. Dec. 25, 1879. Aneroid A .... 21*700 21*700 21-700 21*600 B . . . . 21*170 20*960 20*940 20*870 D . . . . 21*460 21*430 21-450 21*390 E . . . . 21*500 21*500 21*500 21*440 F . . . . 21*220 21*030 20*950 20-800 G . . . . 21*400 21-380 21*300 21*300 Extreme differences 0*530 0*740 0*760 0*800 Mean of aneroids . . 21*408 21-333 21*321 21-233 Mean Merc. bar. . . 21*928 21*912 21*934 21*888 Mean errors of aneroids -0*520 -0*579 -0*613 -0*655 6. Upon December 26, 1879, we encamped on the Arenal (14,375 feet), at the foot of Chimborazo, and 011 the morning of the 27th the mean error of the aneroids was found to be - 0*737, and their greatest difference 0-880. 7. We then moved up to the Second Camp on Chimborazo (16,664 feet), the mean error continuing to rise, and amounting upon December 30 to -0*903. 8. Upon arrival at the Third Camp (17,285 feet) I found that com- parisons at greater heights would have to be made between five aneroids 1 As aneroid 580 was left at Guayaquil, the comparisons are now between the seven remaining instruments. 2 Aneroid C was lost or stolen shortly before arrival at Guaranda, thus reducing the number under comparison to six. 6 COMPARISONS IN THE FIELD. only, as the error which B had acquired was so large that we had already got beyond its range. The mean error of the aneroids at this point amounted to -0-974, and their greatest difference to 1-120 inches. 9. The aneroids D and E were alone taken to the summit of Chim- borazo on the first ascent, January 4, 1880, and these two were taken because they were working better than the others. The readings on the summit are instructive. Merc. bar. No. 558 . Aneroid D do. E 14-110 inches. 13-050 12-900 (by estimation). The mean of the two aneroids is seen to be 12-975 inches, and the error of this upon the mercurial 1-135 inches. I defer comment to a later point. 10. Their prolonged residence upon Chimborazo seriously affected the constitutions of aneroids F and G. The index of the latter instrument became immovable, and the former was afflicted with a quivering action which set observation at defiance. Comparisons for the remainder of the journey were thus restricted to A, B, D, E only, and they are given in the tables upon pp. 10, 11 so far as is necessary to support the statements, a. that the aneroids acquired considerable errors ; b. that they differed amongst each other to a very large extent ; and c. that their means were far from the truth. 11. After we had been three to four weeks in the interior, the aneroids A, B, D, E were found to hold pretty constantly together (or, speaking more correctly, their movements were harmonious), and they seemed to have acquired their maximum errors for the pressures at which they were used. 1 Of the above four instruments B had the largest index-error, and the following table shows that it remained tolerably constant. It then became interesting Date. Place of observation. Merc. bar. 558. Aneroid B. Error of B. Jan. 14, 1880 Chuquipoquio 19 -683 inch 17-820 inch - 1-863 inch Feb. 8, Hac. cle la Rosario 20-805 19-100 -1-705 ) Illiniza (S. side) . 17-239 15-400 -1-839 16, On Cotopaxi . . 17-431 15-650 -1-781 26, Machachi . . . 21-142 19-360 -1-782 Mar. 28, Hac. Guachala . 21-618 19-950 -1-668 May 1 5, Quito .... 21-631 ., 19-990 -1-641 June 8, Illiniza (N. side) . 17-222 15-400 -1-822 to observe whether aneroids which had acquired such large index-errors could be usefully employed for the determination of differences of level. 1 See the last column of the table at p. 11. COMPARISONS IN THE FIELD. 7 12. Upon the journey to the north of Quito I carried mere. bar. No. 558 and aneroids A and B, and upon arrival at the top of the great ravine of Guallabamba took simultaneous observations of the mercurial and the aneroids. At the bottom of the ravine, two hours and a half later, readings of all three were repeated with the following result : Date. Barometer. Read at top. Read at bottom. Mar. 27, 1880 Merc. bar. 558 (red. to 32 Faht.) 21-692 23-929 do. Aneroid A 21-140 23-400 do. do. B 19-940 22-200 The rise of the Mercurial Barometer was 2*237 inches, do. do. Aneroid A ,,2-260 do. do. B 2-260 13. The foregoing experiment is a descending one employing two ane- roids, and the next is an ascending one in which three were observed. Upon the occasion of the attempt to ascend Illiniza from the north, I read the three aneroids A, B, D before departure from Machachi (9839 feet), and did the same at our camp (15,446 feet). Date. June 8, 1880 do. do. do. The fall of the Mercurial Barometer was 3*954 inches. The mean fall of the three aneroids was 3*960 14. Upon the second ascent of Chimborazo I carried aneroids A and E to the summit, reading them at the fifth camp and at the top. Aneroid A became much out of range, and I therefore cannot give its reading. Date. Barometer. Fifth Camp (4 a.m.) Summit (2 p.m.) July 3, 1880 Merc. bar. 558 16-931 inches 14-044 inches, do. Aneroid E 16*060 12*990 The fall of the Mercurial Barometer was 2*887 inches, do. Aneroid E 3*070 15. The examples which are quoted in 12, 13, 14 give the closest coincidences that I can mention from amongst experiments of this order. Upon the whole, it appeared to me that better values could be obtained from aneroids by taking the mean of ascending and descending observations, 1 than by taking the means of either ascending or descending ones alone, and I now give an example in which this method of treatment was adopted. On March 19, 1880, I carried the aneroids A, B, D, E from our lodging 1 When ascent and descent are only a short space of time apart. Barometer. Merc. bar. 558 Aneroid A do. B do. D At Machachi (9 a.m.) 21-176 inches 20-650 19-530 20-290 At Camp (6 p.m.) 17*222 inches. 16*810 15-400 16-380 8 COMPARISONS IN THE FIELD. at Quito to the top of the hill called the Panecillo, on this occasion reading the 'scales of feet' upon them before departure, and again upon return to Quito. From the means of the ascending and descending readings, the summit of the Panecillo appears to be 651*25 feet above the level of the principal Plaza of Quito, which from the mean of twenty-two observations of mercurial barometer by myself is found to be 9343 '3 feet above the sea. I have no observation of mercurial barometer on the Panecillo ; and, if I had, should still quote by preference the independent observations of Messrs. Reiss and Stiibel, who, from the mean of a large number of observations of mercurial barometer, give for the height of Quito 9350 feet, and for the Panecillo (two observations of m.b.) 10,007 feet. Their difference of level therefore is 657 feet, or 5 feet 9 inches more than the height indicated by the aneroids. Barometer. At Quito, in Hotel. Summit of Panecillo. a rise of On return to Quito. = a fall of Aneroid A 10,680 feet 11,325 feet 645 feet 10,760 feet 565 feet do. B 12,310 13,050 740 12,390 660 do. D 11,260 11,950 690 11,340 610 do. E 11,000 11,680 680 11,060 620 Mean of ascending readings 688*75 feet. Mean of descending do. 613 '75 feet. Mean of ascending and descending 651-25 feet. 16. As the journey approached its termination, I became curious to observe how the aneroids would read against the mercurials upon return to the level of the sea. We arrived at Guayaquil again on July 13, 1880, and the barometers were compared against each other from the 16th to the 27th. 1 The error of aneroid A upon the 16th was - 0-361 of an inch, and of E - 0-321, but by the 27th their respective errors diminished to - 0-341 and - 0-291. I have not allowed the index of either to be altered. They continued to recover in the course of time ; and I found, upon January 9, 1885, that aneroid E possessed an index-error of +0*160, which was very nearly its error upon the last comparison in 1879 before our start, when it was seen to be +0*182. Aneroid A did not recover with the same rapidity. Upon January 9, 1885, its error had diminished to 0*200, and in five years more it recovered another tenth of an inch, 1 7. In the tables at pp. 1 2, 1 3 the comparisons of the aneroids A and E are presented separated from the others. These two instruments were those which were most consistent in their behaviour, and were those which were most frequently employed. From inspection of the tables it will be immedi- ately apparent that ' a good return ' is of little value as a test of working. 1 See the tables at pp. 12, 13 for this and for the succeeding paragraph. COMPARISONS IN THE FIELD. 9 Upon the last comparison before departure, these two aneroids possessed almost exactly similar index-errors ( + 1 7 2 and + 1 8 2), and upon return to Guaya- quil their index-errors were not far apart ( - 0-341 and 0-291). It would have appeared legitimate to conclude that their working had closely corre- sponded, but inspection of the last two columns of the table shows that such a conclusion would have been extremely erroneous. The case of E, taken by itself, is still stronger. This, in course of time, ' returned ' almost perfectly ; and inasmuch as this instrument (like all the others) was tested before de- parture, inch by inch, against the mercurial barometer under the air-pump, and corresponded almost perfectly, it would have seemed right to conclude that its readings in the interim must have been nearly free from error. Yet this instrument, at the greatest height at which it was compared, was found to possess a minus error of an inch and a fifth, the value of which, at the elevation in question, exceeds two thousand feet (see 9). 18. Some of the more important conclusions which must be arrived at from consideration of the results of these comparisons of the aneroid against the mercurial barometer are so obvious that I consider it unnecessary even to point them out ; and, in the remarks which follow, I endeavour more to indicate the ways in which the aneroid may be advantageously used, than to emphasize the objections which might be urged against its employment. A. It seems possible, without reference to a standard, by intercomparison of a number of aneroids, to discriminate between them, and to select those in which most confidence should be placed. B. That, with aneroids of the present construction, it is unlikely that decent approximations to the truth will be obtained at low pressures, even when employing a large number of instruments. The errors of the whole series (A - G) were invariably minus ones, and in the worst cases amounted to as much as two inches upon the mercurial barometer. C. That differences of level at great heights (low pressures) may be determined with considerable accuracy with aneroids, even when they have acquired very large index-errors. D. That in observations of this description a nearer approach to the truth is generally obtained by employing the mean of ascending and descend- ing readings than by taking ascending or descending readings separately. E. That the test which is commonly applied of comparing for brief periods (minutes or hours) aneroids against mercurial barometers under the air-pump is of little or no value in determining the errors which will appear in aneroids used at low pressures for long periods (weeks or months). F. That, similarly, comparisons of aneroids against mercurial barometers in balloon for a brief space of time afford little or no clue to the errors which will be exhibited by the former when subjected to low pressures for C 10 COMPARISONS IN THE FIELD. % . o ^ ; ^ p P P s P P P S 8 p |2 00 05 CD o 10 on ! O5 CO ^ CO ^ ^ O5 CO CO o OD 00 I 1 CO CO J-^. gi i i O . . o O CO CO 00 CO o o + 1 1 1 1 1 1 1 1 1 1 1 1 1 1 A ,4 o *t "So .s . . ** ~ *^ 00 *o CO ^f iO ^ 00 05 00 CM ^ on CM CO CO *"* ^ O CD 00 J^*. ^4 fM fM i i CO 00 00 S ^"^ O5 00 oo 00 00 CO O5 05 05 oo 00 ^ "i O5 CM O5 CM O5 CM C5 CM O5 CM OS CM 00 CM CM CM rH CM t i CM 1 1 CM l-H CD i i CO 1 1 OS 3 O .s ~ * *^ R * R " *^ *^ *^ R R *^ 5 CO CO CO (M 00 Q co CM CO CO i I E O5 00 CO CO 10 CM O5 CO CO CM rH CO CO O5 CM O5 CM O5 CM o CM O5 CM CM OS CM oo CM CM CM CM CM CM CM rH 10 14 IQ rH J J Se S 'S ^ Js " >o O O CD iO O CM CO 00 00 CM CO O i i CO CO CO CO 00 CO o w o J~ CD >o J>- 1> oo 00 Jr 00 o O O O O O o o O O f- 00 00 oo oo oo 00 00 * *" * CO CO CO CD CD CO CO 8 c3 o Pi ^ d 3 ii rrt o d 6 1 1 -IJ 3 1-5 ,*? 1 Kingston, f 3 Guayaquil, E Muiiapamba, I J 1 S 1 Guaranda, d P d P d P 1 d P d P O5 00 *"* ** *^ 1 - oo" cT CM" CM" co" J^ oo" ~ co" 10" ^ co" of ft CM rH CM i i r ~ | 1 1 r- 1 CM CM CM CM CM CM k - . *^ O CJ I-H O R ^ p P R ~ - - R - 8 R R R R 6 I-H CM CO Tj4 o CO t- 00 05 O T 1 CM CO Tt* 1C CD J^ COMPARISONS IN THE FIELD. 11 1 II si d CO O5 "# r- O5 o CO CO OS o iO 05 O CM CM O5 CD O5 O5 to JD^ 05 CM CM i l O5 O5 CO ^ O5 o CM 10 i t O5 CM i I O5 CO O5 iO CM O r 1 1 rH I I 1 1 1 1 1 1 1 1 1 1 1 1 1 i-H 1 i 1 1 i-H 1 O 1 i-H 1 cr ^ H Ife d o 00 J* 05 r ( 1 1 I 1 00 CO -tf r-~ CO ^ O5 O JT^ i i i 1 10 00 r- O5 CO CM i l CO * 1 1 CM i I 10 CD 10 CM O5 CD Oi CM O5 T^ CM CO CM CM CM - m ^ CD CO T(H CO O5 CM rH J^ ' "* 1 ' CM ' ' 1 3 4 .s r^ iO IO CM (M rH J^ 10 CO 10 1^ CM iO o O5 CO o i-H i 1 00 CM TH iO iO co CD i-H I 10 10 CM 10 00 rH O5 CD CD CM CD M ' 1 ' 1 ' CM reatest differ- 3nce between Aneroids. 1 to CM 00 GO O 10 00 o CD o CM Tt< CO i i CO p O5 r I iO r^ r- 1 p CM o CM R 00 O I-H ^ 1-H CM No. of G Aneroids observed. QD ^h CM TF T}H * Tt< T^ * T^ "* ^ CM CM ^ CO Tf Place of Observation. Chimborazo, Second Camp H3 J g p * 4-2 1 p 02 A &* a 3 n a 02 & Tambo of Chuqnipoquio . Ambato .... Machachi ... Hacienda de la Rosario 0) g 02 I J '3 i i 4 1 Machachi . , . 3 s Quebrada of Guallabamba (top) 3 1 6 <-G 6 ft .- O 1 1 02 ti P! PH N 1 1 1 Machachi . O5 j>. CO 00 rH 00 I 1 R R 8 R R * 1 CO s iO 4< ^ O5 ^ r- 1 CM CM O5 CM 00 ^ O5 CO I 1 CD CM CM r^ CM j 10 i-H | co 09 i i rH ^ O J2i 00 I 1 O5 r- 1 CM r- ( . CM 00 CM O5 CM o CO i-H CO CM CO CO CO ^ CO 12 COMPARISONS IN THE FIELD. ,* id 5 d (M 00 rH Tt< O i> rH rH t^ 1> co 00 CO rf CO ^ C 00 m + 1 1 O 1 o 1 I 1 1 o 1 o 1 o 1 i i 1 1 1 O | 1 o 1 1 ^ < o I CM !> i i ^ "* o i> r- 1 CD X O J^ co I 1 oo CO CM o CO CM t^ CM CO rp ^ o >o -^ 00 >o oo oo o CO o o JO ^ ^ CO *G IO o 1C O5 O5 "* i + 1 1 o 1 1 1 1 1 1 1 1 1 1 o 1 1 o 1 I 4 LU a a H3 O5 o 00 o t^ o o c o * o 10 ^ "tf CN CO CD r^ J^ o CD CM o O5 o CO "tf >o t^ ir- o o ""* CM CO o O co CO < 05 05 OS GC I 1 I 1 r^ m >o >0 CM >o 00 pH c CD GS GN GN GS *" pH ^ < !2 a .s O "* O5 1 1 oo r~ JO oq p o rj< CO i^ O CO o CD CO >0 l> 05 >o CO Tj< 00 00 >o CM i i o -f CD 00 XS o w ^* r^ 4 O5 O5 05 00 1 1 1 1 r^ CO >o >o >o O5 pH 6 CO GN GN vl rH O oo -w ^ - js5 *o* |j li"3 i rH 00 CO t^ Tj< >0 00 r^ so t~ rH i i CO t^ !> t^ 00 CO 05 1 1 CM t^ 00 CO i> ^ 00 "t 05 i ( i i r I 00 CD Tt< CO 00 CD i i Q Tj< 1 1 *Q ID X O5 CO CM |1{ O5 05 O5 ob r-i rH r^ CD CO CO "* CO 05 oq I 1 O r- "^ r ^ Place of Observation. London . Colon, Isthmus of Panama Guayaquil, Ecuador . 0* rr 1 cf E Tambo Gobierno, Ecuador T3 1 O nd 0* ft Chimborazo, First Camp 3 13 g ra a -1 -1 & ni 1 & 8 -^> 02 & PH <-3 $ 6 n Tambo of Chuquipoquio Ambato . Machachi . Q oB i B -r | o 'i s CO 1 1 *n CM oq 00 CM ^ o CO >o TJ< 1-5 ** O5 "^ i i fM 'Tl fi O5 CM GO 1 O5 ^ rH GO O5 o i 1 CM CO ^ id CO * 00 COMPARISONS IN THE FIELD. 13 -* H Q R d r 1 CO 00 CM 00 r~ r- r- ^ Jr^ CM CO r- 10 oo ^ F 1 00 i I r^ 00 TjH to - . rH CM CO I 1 O5 CM O CD rH i 1 1 1 1 o 1 1 1 1 1 1 1 o 1 1 + 4 < o I 1 i i r^ <* CM Oi tO J>- r^ o i i CD O CM CD O CM i i ^ I 1 o O5 J> Tt< 1 1 CO CO . r-H CD CO i i -# CO O CM i 1 O 1 o 1 1 I 1 1 1 1 O 1 o 1 O 1 r3 ui s d r-t CD Oi CM o 00 C5 oo 00 00 o Oi CD o J^ o CO o o Oi Oi O 00 CO O CM o t^ ^ o rH Oi o O5 to to CO o Oi fj CD i I O CM CM CM CM r^ r i to 1 1 CO I i CM r 1 Oi rH i i CM rt< CM t^ (M Oi CM Oi CM Oi CM rC! 8. < ]g p i CD A 10 to o JC^ 1 1 00 CO tO CO co nd its rai o CO tO CM tO oo CO Oi tO o o CD "* to ^ CD r 1 CM CM CM I 1 CM CO r-H 00 I 1 CO 1 ( CD I 1 $*, M Oi I I i 1 CM "* CM 1^ CM Oi CM Oi CM Oi CM 2S?J .ja TO iss d o S3 t i CO "* CM <* r 1 ~ J O . I CO CD CM CD CM CM CM o TH -tf 1 1 to rH CO Oi TP -^ o . . . rH i 1 Oi i 1 Tj< Ci ^ J^ Is? JC^ r I i 1 rH r 1 J>- 00 CD CD ^ Oi Oi Oi r ~^ C*M Place of Observation. g PH 1 f O O 1 3 O 3 ft 0* ft 0> ' U 'a ti ,p 1 a 3 i Carihuairazo, Camp on 4^ 2 2 P3 02 ft Chimborazo, Fifth Camp M O -4-3 s pj Ul & Commencement of descent towards Pacific Ocean 1 & .s Pn O Hacienda of Cayandeli Camp near Bridge of Chimbc Guayaquil ft London o 00 ^~- to 00 00 r- 1 00 r 1 CD i i 4 CD CM CM o I 1 ^> CM 00 0) s 00 CM OJ CM CO X * o rH S i i rl CO i i 1^ CM OS d A ^ ^ ps 1-5 & 1-5 c3 -9 o SQ OS 1 1 O CM i l CM CM CM CO CM TH CM 10 CM CD CM J^ CM 00 CM O5 CM O CO rH CO CM CO CO CO TjH CO 10 co 14 COMPARISONS IN THE FIELD. prolonged periods. [The balloon test is only a repetition of the air-pump test. In the former case the instruments are exposed to a natural, and in the latter case to an artificial diminution of pressure ; and if the duration of time is equal in each case the results ought to correspond exactly.] G. That very material errors may be fallen into by regarding 'a good return ' at the level of the sea as a proof of correct working, at low pressures, of aneroids of the present construction. H. That for the detection of such errors as aneroids (of the present con- struction) will exhibit when subjected to low pressures for a length of time, aneroids should be subjected artificially to similar pressures for a long period. PART 2.-EXPERIMENTS IN THE WORKSHOP. Throughout the second part of this paper, as in the first part, my remarks are principally confined to comparisons of the aneroid against the mercurial barometer. I enter only incidentally into consideration of uses to which these instruments can be put, and do not attempt to explain the mechanical im- perfections that cause the errors to which reference is made. The question which I attempt to answer is, Do aneroids, under ordinary conditions, read truly against the mercurial barometer ? The earliest experiments which were made in the workshop were under- taken with the view of confirming or upsetting the conclusions which were arrived at from observation in the field ( 18), and they were directed first of all to learn whether it is a fact that all aneroids lose upon the mercurial barometer * if submitted to diminished pressure for a length of time ; to observe the length of time during which the loss continues to augment, and to ascertain the extent of the loss that occurs. These earliest experiments were made entirely with aneroids of Mr. J. J. Hicks' manufacture, and they confirmed the observations made in the field. I felt, however, that it was desirable to include instruments by other makers, and I continued to experiment until 70 aneroids had been submitted to ex- amination, and could be reported upon. 2 Every aneroid which was tested, without exception, lost upon the mer- curial barometer when submitted to diminished pressure for a length of time that is to say, for a day, a week, a month and upwards. It was found that the greater part of such loss as occurred took place during the first week. The experiments were then continued to endeavour to learn what proportion of the loss which occurred during the first week took place during the first day. When this class of experiment had been continued for many months, I turned my attention to the behaviour of aneroids upon their being allowed to return to normal pressure after they had experienced diminished pressure for a length of time. It was found that they always recovered a large part of their loss, and sometimes gained more than they had previously lost. It was found that the recovery might extend over several weeks, and that the greater part of the recovery or gain usually occurred in the first week. As the amount which was recovered very seldom exactly equalled the amount of the previous loss, in the great majority of cases there was a marked change in the index-errors, and, in some instances, a large alteration. 1 See 21 for explanation of the expression "loss upon the mercurial barometer.'' 2 A number of others were tested. Some behaved well, but they are not included here, as they did not bear makers' names. 16 EXPERIMENTS IN THE WORKSHOP. I then endeavoured to learn whether aneroids which possessed large index-errors could nevertheless be usefully employed for the measurement of variations of pressure, and found that in the majority of cases they might be so employed. Other experiments having a practical bearing suggested them- selves from time to time. 19. The method by which the aneroids were tested in these experi- ments was that which is habitually employed during * verification ' of these instruments. The aneroids were placed under the receiver of an air-pump, to which there was an attached mercurial barometer. Upon air being with- drawn, there was simultaneous reduction in pressure both for the aneroids and for the mercurial The indications of the latter were frequently checked by reference to a standard mercurial barometer, hanging alongside. 20. Although my method of testing was the same as that employed during ' verification,' there was one essential point of difference between the verification tests and my own, namely in the length of time during which the aneroids were kept at reduced pressures. I have good authority for say- ing that even when an aneroid is verified at Kew Observatory inch by inch, down to as low as 15 inches, it is unusual to. occupy more than an hour in the operation, about one-half of which amount of time will be consumed whilst pressure is being reduced, and the rest while pressure is being restored. This, also, is about the length of time ordinarily occupied during the manufacture of aneroids upon the process termed 'pointing, 5 i.e. laying off the scale on the dial of an aneroid (by comparison with an attached mercurial barometer) prior to graduation. Therefore, what verification amounts to is this. It is a repetition of the temporary reduction of pressure to which aneroids have been subjected in the course of manufacture ; and one learns from verification whether the ' pointing ' and the subsequent graduation have been accurately performed. One does not learn from it the errors that will be manifested by aneroids which may be subjected to a reduction of pressure for a greater length of time. In order that my tests in the workshop might be the equivalent of the tests to which aneroids are put in the field I kept the instruments at various pressures between 26 and 14 inches for periods of days, weeks, and even months at a time. 21. The first six aneroids that were experimented upon were taken from the stock of Mr. Hicks, and were marked temporarily 1-6. These were placed under the receiver, and had pressure reduced to 22*5 inches, and were kept continuously at that pressure for six weeks. The annexed table scarcely requires explanation. In Column 1 the errors of the aneroids are given which were exhibited upon their being reduced to 22 -5 inches. In Columns 2-7 the errors are given which they showed at the end of each week, during six successive weeks ; and in Column 8 the total amount of the loss is stated that occurred during the six weeks. In the case of aneroid No. 1, a + error of 0-159 was converted into a - error of 0-335. The actual loss amounted therefore to 0-494 of an inch. In the case of No. 5, the error of 0-016 was increased to -0-485. The actual loss was therefore 0-469 of an inch. The amounts so lost I term " loss upon the mercurial barometer." EXPERIMENTS IN THE WORKSHOP. 18 EXPERIMENTS IN THE WORKSHOP. Several points came out very clearly during this experiment The first and principal one was that the whole six instruments lost considerably upon the mercurial barometer. Another, and it seemed to me an important one, was that the greater part of the loss occurred in each instance during the first week In every succeeding experiment these facts were confirmed. Another point was the large difference in the loss in the different instruments the least being 0-209 and the greatest 0*494 of an inch. The gradual stoppage in increase of the loss is best seen by examining the line in which the mean errors of the whole are given. At the end of the fifth week there was no perceptible increase in the mean error, but at the end of the sixth week a slight increase appeared. This was due to Nos. 3 and 5, which probably would have continued to lose some small amounts for several weeks longer. 22. The emphatic manner in which this first experiment in the work- shop confirmed the experiences in the field keenly interested those with whom I was associated, and immediately upon its termination we started a fresh series of aneroids (marked temporarily 10-12) at a pressure of 17 inches, and kept them at that pressure during five weeks, with the following results. EXPERIMENT IN WHICH THREE ANEROIDS (No. 10, WATCH-SIZE ; Nos. 11, 12, THREE INCHES DIAMETER) WERE KEPT AT A PRESSURE OF 17 INCHES FOR FIVE WEEKS. Aneroid. 1. Errors of aneroids at 17 inches at start of the experiment. 2. Errors of aneroids at the end of the first week. 3. Errors of aneroids at the end of the fifth week. 4. Loss of aneroids upon the Merc. Bar. in five weeks. inch. inch. inch. inch. No. 10 (Hicks). -0-023 -0-731 -0-944 0-921 ,, 11 , -0-203 -0-721 -0-909 0-706 12 -0-043 -0-531 -0-709 0-666 Mean errors of aneroids ^ on Mercurial Baro- 1 meter (reduced to 32 f Faht.) ) -0-090 -0-661 -0-854 0-764 Note. In calculations for altitude, the value of 0'764 of an inch, at a pressure of 17 inches, is about 1220 feet. I do not feel it necessary to give this second experiment in as full detail as the previous one. Each aneroid lost considerably upon the mercurial barometer ; in each case the greater part of the loss occurred in the first week ; the loss was different in each instrument ; and the loss seemed to cease to augment about the fourth or fifth week. No. 10 lost nearly an inch (0-921), and the mean loss of the three instruments amounted to 0-764 of an inch. EXPERIMENTS IN THE WORKSHOP. 19 23. When this experiment was concluded I started a third series of aneroids (Nos. 7, 8, 9), three inches diameter each, at a pressure of 1 6 inches, and kept them at that pressure for two months. This series lost less at 16 inches during 'six weeks than Nos. 10, 11, and 12 had lost at 17 inches during five weeks. But the loss, as before, was considerable in each instru- ment ; the greater part of the loss occurred during the first week ; and the loss was comparatively trifling after the fourth week. EXPERIMENT IN WHICH THREE ANEROIDS (EACH THREE INCHES DIAMETER) WERE KEPT AT A PRESSURE OF 16 INCHES DURING EIGHT WEEKS. Aneroid. 1. Errors of aneroids at 16 inches at start of the experiment. 2. Errors of aneroids at the end of the first week. 3. Errors of aneroids at the end of the second week. 4. Errors of aneroids at the end of the third week. inch. inch. inch. inch. No. 7 (Hicks) + 0-104 -0-118 -0-180 -0-206 55 & 55 + 0-069 -0-223 -0-305 -0-306 9 . + 0-014 -0-368 -0-405 -0-451 Mean errors of ane-^ roids on Mercurial I Barometer (reduced f to 32 Faht.) J + 0-062 -0-236 -0-297 -0-321 5. 6. 7. 8. 9. 10. Errors of aneroids at the end of jhe fourth week. Errors of aneroids at the end of the fifth week. Errors of aneroids at the end of the sixth week. Errors of aneroids at the end of the seventh week. Errors of aneroids at the end of the eighth week. Loss of aneroids upon the Merc. Bar. in eight weeks. inch. inch. inch. inch. inch. inch. -0-252 -0-276 -0-285 -0-306 -0-271 0-375 -0-322 -0-326 -0-355 -0-331 -0-336 0-405 -0-472 -0-476 -0-495 -0-506 -0-531 0-545 -0-349 -0-359 -0-378 -0-381 -0-379 0-442 Note. In calculations for altitude, the value of 0'442 of an inch, at a pressure of 16 inches, is about 750 feet. 24. After this experiment, I proceeded to test four aneroids which had been specially made, and embraced some peculiarities in construction. In this (the fourth) series each instrument lost considerably upon the mercurial 20 EXPERIMENTS IN THE WORKSHOP. barometer ; the greater part of the loss occurred during the first week ; and the mean error of the whole four ceased to augment after the fourth week. EXPERIMENT IN WHICH FOUR ANEROIDS (EACH FOUR AND A HALF INCHES DIAMETER) WERE KEPT AT A PRESSURE OF 19 INCHES DURING FIVE WEEKS. Aneroids 13, 14, 15, 16 (Hicks). Mean error of the aneroids at 19 inches at start of the experiment. Mean error of the aneroids at the end of the first week. Mean error of the aneroids at the end of the fifth week. Mean loss upon Mercurial Barometer in five weeks. inch. inch. inch. inch. Mean errors of the ftnir^j aneroids on the Mer- 1 curial Barometer (re- | ducedto32Faht.) J + 0-013 -0-246 -0-348 0-361 Note. In calculations for altitude, the value of 0'361 of an inch, at a pressure of 19 inches, is about 520 feet. 25. These four experiments (as well as subsequent ones) showed clearly that the greater part of the loss which occurred took place during the first week. From the following table it will be seen that in every series the mean loss in a week exceeded two-thirds of the total mean loss upon the mercurial barometer. Number of Aneroids employed. Pressures at which they were kept. Length of time during which they were kept at these pressures. Mean loss of each series upon the Merc. Bar. in one week. Total mean loss of each series upon the Mercurial Baro- meter. inches. inch. inch. 6 22-5 6 weeks 0-298 0-373 3 17 5 0-571 0-764 3 16 8 0-298 0-442 4 19 5 0-259 0-361 26. As it appeared from these experiments that the greater (and more important) part of the loss which occurred took place in the first week, I con- tinued by testing aneroids, for one week each, at every inch of the barometer between 14 and 26 inches, and endeavoured to procure instruments by a diversity of makers and of various diameters. This series of experiments necessarily extended over several years, as it was seldom possible to test more than a few instruments at a time, owing to difficulty in obtaining at any one time a number of instruments with similar ranges. It included aneroids by Hicks, Casella, Adie, Elliott, Negretti and Zambra, Cooke and Sons, Secretan, and Hilger. EXPERIMENTS IN THE WORKSHOP. 21 'S PI 15?! oo ^ r- co co JC 1 "* CO C5^l IH iO G^l G^ C D O rH CM Tj< "i ^ O5 O5 CO Jt CM COtoO5OOCMlr-J: - co to co j> PH ai"* gill -* CO CM O O tOTt*iOtOiOO*O.t - O CO CO CO sj!s ^ 1 O5 00 CO lr- rH CM^lr^OOOOO^C O t- CO Tf to D rH rH rH 02 03 2 C J" s 6 O O O O O C D "fa O CO J>-OOO5OrHCMCO^ to CO Jr- 00 22 EXPERIMENTS IN THE WORKSHOP. fill -iJ 1> i 1 00 Ol CO iO * 01 Ol 01 I 1 CO CO O5 CO lO "^ ^^ "^ CD *O ^O CO 00 O I-H GO O O CO CO CO CO I-H CO CO O 01 JL- I-H g 01 Ol Ol Ol 01 Ol CO 01 01 rH Ol Ol CO CO '^f CO "^f 0*00 .= 666 o o o o o o o 00000 5^.S III . j; ~ 10 * * 2? ^* " Ol ~ 00 ~ '^f l&* - 1 CD CO |*|| , - CO - 01 r. ^ Ol Ifij 1 S " Ol " Ol .11? Hi 1 , 3? 01 a- CO i-tN MHf Ol Ol Ol Ol 01 5-s . . g o ffl o ^H ^^ JH Property I i e3 O S o* o 73 o 1 I EH fl OJ rt O O O -^ g o W Q) rf H 8. a >> p s J> 1 10 1 oT O5 CD O O CO "* CO Ol O I-H CO 1 v_x v_x 00 2 J-j I 1 1 1 ' ' 'I ' ' ' -d T$ ^ 1 0. fl ^4 03 o .,_, i M C3 a So? w f'** tf d o d W d d o 6 o M 5 HH "~^ *"^ '^ *?? gv- h-j WD 1-5 *~d ^ !- W ^-J- 1 t-5 ^ -a & Ha EH Ha .4 W O5 O I-H ^jq CO Tf iO CD Jr^ 00 O5 O I-H Ol CO ^< iO CO 01 Ol Ol 01 Ol Ol Ol CO CO CO CO CO CO CO EXPERIMENTS IN THE WORKSHOP. 23 I'aljS g ff - rH CO 1^* CO OS * J^* CM 00 CO 00 l>* - r 1 rH co os tr^- OS Hgl COOrHCOi^l>-CMCMl.^^CMOOOiOCO Sr^CSCOCOOOCOCOiOi^CO^CMOSOCMCO g^OT^iOCMCOCOCOCOi^OOr-lCMCOCO^ ^ co r-~ 10 ^ CD SjJ.s .Soooooooocoooocboo O O S"S D o o o o o o ^ O 00 CM O iO CM ill jfa ^ "^ ^^ ^" *^ ^* ^ J> ?; ^ ?; p P OS ^ 00 OS OS IH CM CO rH i 1 rH |||| iO 2 CM i i i '^.O^^^^^^OS 00 CQ rH ^ ^ oCMCMCM *^CM ^^^^^N^,.., *> A ^ *N N I|P .s 1 s s (5*o "" g 'oCMCMCMCMCMCMCMCMCMCMCM'^T^T^^Qq CM CO feH 1 1" S OJ 'S 5* ^ ,2 ^ ^^^00000006000 O . O i * 1 r^H 13 r^ H W H g ^^ CO U-O^H(MOOOCMt-CMOS^co X iO N C*rcO X OS > ^OOCM'^} < Tt l COOSCOOSCOil r _ii ii iO 10 00 00 OS iO ^-^ i 1 iO iO O M* iO I 1 3 fl s 1 t I O i I O r5 s sSs^W^^gW^WcSdddS O t-i O w i-i OH^QI-; Q ^ ^ h^t-51-^ H-5 M :*"> 1-4 *TJ 1-4 o fe t" OOOSOi iCM'cOTl*io'cdl>-odO5di-HCM co n! xO ^O 24 EXPERIMENTS IN THE WORKSHOP. BKI P,+.C O5OC5CiCOCO.t 00 t^OOCMOCOCDCMr^ lOr-OOOOrHrHOCO I I rH CM CO CM 00 ^ CD Oi iO O^ CD !> iO iO CD CO O CM CO iO *x s CM CM 00 00 lo CO CO ^f 66666 J>-OOI>-l>-CMCMCMiOiOrHCDCOCO COOCMCDCMOiOOOCMO^OCO Ol>-r~CMCMCNCOTtr^aOO OrHOOOOOOOOrH II * B s 3 , rH (M , g tc 1*5 I r- ^^^^^^^co -^^^10 ^^ rH ~rH ^ I -H'~rH' s CMCMCOCOCMCMCMCMCOCOCOCOCMCOCOCMCM s & - & ' ' ' i ' O a) OD Oaj OJ QJ ^ oo- ^ o H o g ? ^ j^d -3 a 4 -a s 4 -a a -I 4 I - ^ J ? " ^ ^ -j > I > - HI EXPERIMENTS IN THE WORKSHOP. 25 These tables speak for themselves. I draw attention, however, to the last column of each, in which there is shown approximately the errors in determinations of altitude which would have been probable if these particular aneroids had been employed in the field, at the mentioned pressures, for one week ; and I point out especially that these errors, serious as they are, are (taking them as a whole) probably only about two-thirds of the maximum errors which these particular aneroids would have developed, if they had been kept continuously at the mentioned pressures for one month and upwards. 27. It appears to me to follow, and to be indubitable, that a great part of the altitudes throughout the world which depend upon observations of aneroid barometers made while ascending must be too high, and that a general lowering of them will be found necessary ; and that it is probable the reduc- tion in height will have to be at a greater rate per cent in the case of the loftier positions than in the case of the inferior ones, and with those which are in the interiors of continents than with those in the neighbourhood of the sea-level. 28. Concurrently with these observations to attempt to learn the extent of the loss of aneroids upon the mercurial barometer on being sub- mitted to diminished pressures for a week, I observed the loss which occurred in the first day of the first week. This is variable, sometimes being about one-third of the week's loss, 1 and sometimes more than three-fourths of it. This is illustrated by the examples which are given on p. 26. 29. The experiments which have been already quoted indicate that the loss in aneroids upon the mercurial barometer augments at a constantly diminishing rate, provided they are kept continuously at the same pressure. The loss is greater in the first week than it is in the second, or in any suc- ceeding week. It is rare, however, to observe in any seven successive days a perfectly regular (or symmetrical) increase in the errors. Minor imper- fections of construction prevent perfectly harmonious readings, and render it impossible to say more than that the loss during the first day of the first week (so far as my observation extends) is almost always greater than during any subsequent one in the first week. A few examples are given upon p. 27 in illustration. 30. The loss in aneroids upon the mercurial barometer can be observed in almost all instruments during the first hour they are subjected to dimin- ished pressure, if the diminution in pressure amounts to several inches ; and in aneroids giving the inch of the mercurial barometer that length upon their scales, and more particularly upon aneroids with expanded scales (Hicks' Watkin Patent) the loss may even be traced in successive hours. 2 The loss in the first hour of the first day, so far as my observation extends, is always greater than in any subsequent one. 1 I have observed some exceptional cases (not quoted in the table) in which the day's loss has been less than a, fifth of the week's loss. 2 When standing over these instruments, whilst they have been kept at so moderate a diminution in pressure as 26 inches, I have seen the index move backwards whilst the mercurial barometer remained immovable. E 26 EXPERIMENTS IN THE WORKSHOP. TABLE SHOWING THE AMOUNTS LOST UPON THE MERCURIAL BAROMETER IN ONE DAY AND IN ONE WEEK. Diameter Pressure at Loss upon the Total loss upon Aneroid. of instru- which it was Mercurial Bar. the Merc. Bar. ment. kept. in one day. in one week. inch. inch. inch. inch. Hicks (E) 2 14 0-510 0-803 do. ... (F) 2 0-720 1-033 do. . . . (22) 3 15 0-546 0-701 do. ... (23) 3 j? 0-481 0-746 Casella . . . (5606) 2 17 0-521 0-727 do. ... (5606) 2 18 0-378 0-463 do. ... (5798) 3 ?) 0-513 0-653 do. . . . (5792) 2 20 0-660 0-767 do. . . (5892) 2 5J 0-545 0-652 Hicks . . . (37) 2 55 0-545 0-677 do. ... (38) 2 55 0-280 0-362 do. . . (39) 2 )> 0-685 0-842 do. ... (40) 2 )) 0-255 0-337 do. ... (42) 2 0-225 0-287 Casella . . . (580) 2 21 0-325 0-461 do. ... (1021) 2f >> 0-385 0-533 do. . . . (5793) 2 22 0-342 0-473 do. . . . (5893) 2 j> 0-227 0-353 Hilger . 2 jj 0-292 0-458 Hicks . . . (20) 4| 23 0-154 0-252 Cooke . . . (160) 2} 0-161 0-264 Hicks (Watkin Patent, 232) 3 24 0-111 0-273 Casella . . . (4493) 4 24-5 0-146 0-185 do. . . . (5682) 3 5) 0-181 0-245 do. . . . (1842) 4j ?) 0-121 0-200 Hicks (Watkin Patent, 17) 4 i 25 0-052 0-144 do. ( do. 9) 4 26 0-069 0-087 do. ( do. 141) 3 >> 0-084 0-094 Casella . . . (2923) 4 ) 0-099 0-132 EXPERIMENTS IN THE WORKSHOP. 27 A C Or * rHrH^OSrHCOrHCDt-OSOS^OSOSCOlOCDrHCDCOCM =5^^ rHOiioos^rHi loot^r^r^t^i looosiococOrHCMCM g W - ^OrHCDT^CaOCMCOlOOlOCOCMrHOrHrHi ICMrHOO p S J5 > .Sr^rHOOOrHOOOrHOOOOOOOOOOO |5i i i i i i i i i i i i i i + i i i i i + i J3 *; " r OOOSOSCDrHrHCDCDCDTHrt<^iOOi01^CMCMCOr- -^ =t! COCDOSOSOOOr^iOiOCOCOCOt^CMCOOSCMOOSCOO g W g ^OpOlpCOOOrHCMlpOlpCOCMOrHrHOrHCMOpp 0^0 JorHOOOrHOOOrHOOOOOOOOOOO Il^s i i i i i i i i i i i i !+ i i i i i + i cS-g ,^,0^ OS-^OO> CO^rHrHCDCOt-OOOSOSOSCMOO "^4?^^ CDCDCD^^ Or-OSOSOi>-rHiOOOSCD^O g^gg ^r-O5lpCO^ ^ ^O^CMrHOrHrHOrHrHOOO gH^ 6 6 o 6 ' ~6rHo6666666666 1125 i i i i +1111 + 11111 + 1 * S ^^^b COOOOS^OSOS^OSOSrHOOCOJ^CMOOt-J^t-JC^CD^ ^^ r=5 ' rHr-OSCOOiOOt^iOCO^^t^OSOSOOSCDCOiOO g m g^ ^'J^CO^CMl^pCM^p^CMrHprHppprHppp g2^1 .SOOOOOrnOOOrHOOOOOOOOOOO IriSs IIIIIIII + III++I + III + I W ? :>Or ~ l 9 o 9'~ l 9 'S o^o be ^-2^ |JoOOOOOOOOrHOOOOOOOOOOO HS-P'S ' I++I I + '+ + I J- 4- -J- J- 4. -I- J- 4. 4. 4- 4- s.-5-s O ~~~rH ^CM CM ^^^CM ~CM *" ^CO ^-^^ ^^CD ~CM "CMCM "''CM CMCOCOCMCMCMCMCMCMCMCMCMCM^CO^CO U_CMCOr~-OSCMCMOrHCOCO OOCMCOCMCMCOOS ^~* CMCMCOCOOSOSOOCMOSOS JCDCMCOOSCO'^'CM ^ ^ ' ^ ' * ^ ^ 1^* 00 iO O JC^* 00 I rH QQ ^ CD 00 O5 ~*^ 10 10 N 'rH 10 10 v -' Tj5 the readings are given which were taken when the mercurial barometer indicated a pressure of 23*929 inches, and in Col. 6 the rise is stated which was indicated by all the barometers (the differences between the readings in Cols. 3 and 5). In Cols. 7 and 8 the readings and errors are given upon return to natural pressure. 1 To have repeated the conditions of the experiment in the field, these aneroids should have been kept at 21*692 inches for four or more weeks. There were other demands upon my apparatus at the time, which prevented me from doing so. 38 UPON THE USE OF THE ANEROID BAROMETER 00 I 1 (MOJOi icoOi i O i < i i i o eo G 0Or-iOOOOOOOOOiO OOiOOOlJ^CDOCOi^-^COt^CO^OOCDCOOiOOO " 25 i-t* ^rS C : c - a; - C N p-s I g (^O^ ^ ooooooocoooooooooooooo 1 : OOOOOOOOOOOr-iOOOOOOOOi-iO 01 w O5 5 > t OOOOOOOOOOOO-OOOOOOOOr-iO + + + +I+ + + + +I I++I 1 1 l-f+l 1 OOOO>OOOOOiOOOOOOOOOOOOO OOOO>OO i iC3Ot^CDO its So r r r r r IN DETERMINATION OF ALTITUDES. 39 The immediate object of this experiment was to see how repetition in the workshop would accord with the experience in the field. It will be found that it agreed in a remarkable manner. The mean of the whole of the aneroid differences in Col. 6 is 2*218 inches * or, if the reading of No. 1021 (Casella) is rejected, 1 and the mean is taken of the twenty-one others, it will be found to be 2*231 inches, a difference of only six- thousandths of an inch from the rise in pressure indicated by the mercurial barometer. If the aneroid observations are taken singly, it will be found that six of them differ from the mercurial barometer only to the extent of 0*018 of an inch or less, and that the extremes (rejecting No. 1021) are 2-130 and 2*360 inches. From this experiment (as well as from many others) it appears that although aneroids under diminished pressure acquire considerable index - errors, and often acquire huge ones under a long continuance of largely diminished pressure combined with frequent fluctuations in pressure, they still continue to march inch by inch with the mercurial barometer ; and upon a single operation involving a difference of pressure even of several inches may come very near to the truth, provided the operation is quickly performed. 47. The correct measurement of differences of pressure is, however, only a step in the barometric determination of differences of level. Every inch of the barometer has its own special value in calculations of alti- tude. The inch embraced between 31 and 30 inches is worth about 894 feet, while that between 15 and 14 inches is worth about 1880 feet. In order to ascertain differences of level correctly it is necessary for an observer to know not only the difference of pressure but also the par- ticular inches embracing his difference of pressure. The value in English feet of the difference between 19*940 and 22*200 inches (the readings of aneroid B at the ravine of Guallabamba) is ten per cent greater than the value of the difference between 21*692 and 23*929, the readings of the mercurial barometer. 2 Even for measurement of differences of level, it is highly desirable that a traveller should, at all times, have the means of ascer- taining the index-errors of his aneroids. 48. The importance of being able to do this is far greater in observa- tions for determination of elevation above the level of the sea. The mistake which the index-error of an aneroid then introduces into calculations is the whole value of the difference of its readings at the upper station from the true barometric reading. 3 Aneroid B, at the top of the ravine of Gualla- 1 As it fairly may be. This is an old aneroid, which, prior to this experiment, had been reposing in a drawer for about eighteen years, and did not work freely. 2 For the purposes of this passage, and for all the computations which have been necessary throughout this paper of the value of the barometric inch at different pressures, I have used the Table prepared at the Royal Observatory, Greenwich, en- titled Corresponding Numbers of Elevation in English Feet, and of Readings of Aneroid or Corrected Barometer in English Inches ; the Mean of Atmospheric Tem- peratures being 50 Fahrenheit. See pp. 56, 57. 3 See the last columns of the tables upon pp. 21-24, 40 UPON THE USE OF THE ANEROID BAROMETER bamba, read 19*940 and the mercurial barometer 21'692 inches. The difference of these two readings is 1-752 inches, and the value of this, at the elevation in question, is above 2000 feet. Although this is a gross case, more extreme ones are possible. 49. I desire to lay especial stress upon the changes which occur in the index-errors of aneroids. From prolonged observation, it appears to me that they seldom remain stationary, and are never permanent. They take place from a variety of causes. In the course of my experiments in the workshop I have seen the index-errors of an aneroid grow to as much as four inches ; in several instances there have been alterations of more than an inch ; and in numerous instruments there have been alterations from scarcely appreciable errors to 4- or errors of two to four-tenths of an inch. 1 Even the least of these amounts (0'200 of an inch) is of consequence in the calculation of alti- tudes, and it has a considerable value when great heights (low pressures) are concerned. The continual variations which occur in the index-errors of aneroids have not, I think, hitherto been sufficiently recognized. Index-errors have gener- ally been treated as if they remained constant ; and, to correct them, it has been considered sufficient to apply as constant corrections (perhaps for the space of years) the amounts stated in " certificates of examination." These certificates 2 have a very limited value. They point out, no doubt correctly, the errors which were exhibited at the time of examination ; and, if the certi- fied aneroids are employed speedily after that date, and again experience reductions in pressure of tlie same amount, and for the same length of time as during the examination, it may be useful to pay attention to the errors which can be deduced from the certificates. A single week, however, of active use (that is to say, of considerable variations in pressure) may considerably alter the index-errors, and will certainly change them to an appreciable extent, while under more prolonged use further variations will most likely occur. It seems to me, therefore, indispensable for all those who aim at correct measurement of altitude above the level of the sea, by means of aneroids, and especially for those engaged upon prolonged journeys, to be at all times able to determine their index-errors. 50. I now draw attention to a feature in the working of aneroids which has as yet only been casually referred to in this paper, namely, the want of correspondence between differences of pressure indicated by them in ascending and in descending observations. Under certain conditions descending readings indicate less difference of pressure than ascending ones, and under other con- ditions the reverse may be the case. These facts can be utilized in deter- 1 The index-errors stated in Col. 2 of the table upon p. 38, with the exception of F (Hicks), 580 and 1021 (Casella), and 170 (Cooke), were all acquired in the course of the experiments which are described in Part 2. One of them, it will be seen, is as much as +0*478, and another -0*407 of an inch. These alterations occurred though the instruments were used with the utmost tenderness. Under the rougher usage to which they would necessarily be submitted in the field there would be liability to still greater changes. 2 In their present form. See pp. 51-54. IN DETERMINATION OF ALTITUDES. 41 mination of altitudes. I shall presently endeavour to show experimentally the effect which is produced by the intervention of a greater or less space of time ; but, before doing so, I will restate a few of my facts in a somewhat different manner. 51. All aneroids commence to lose upon the mercurial barometer directly pressure is reduced. They lose a certain amount if pressure is reduced to the extent of an inch, they lose a larger amount if it is reduced 2 inches, and more and more for each successive inch. Also the loss is greater in a day than it is in an hour, greater in a week than it is in a day, and it continues to augment perceptibly for about a month. Further, immediately pressure commences to be restored, an aneroid endeavours to recover the previous loss. This latter tendency is more and more important as each inch of pressure is restored, and after return to normal pressure it continues to operate (though with constantly diminishing force) for several weeks. 52. It must therefore be apparent that the difference of pressure (or difference of level) which will be indicated by an aneroid between any two stations will be influenced by the greater or less amount of time which may intervene between the readings at the two stations. In the case of a moun- tain 5000 feet high (or any other height) one result will be attained from an aneroid if there is an interval of four hours between a reading at the bottom and another at the top, a different result will be attained if there is an interval of twelve hours, and yet another if there is an interval of twenty- four hours ; and the difference of pressure (or the altitude) which will be indicated will be greater in the second case than in the first, and greater still in the third than in the second. It will not, perhaps, be so readily apparent why descending readings, closely following upon ascending ones, always differ from them, and I will state three imaginary cases in which this will occur. 53. Let us imagine the case of an ascent being made in five hours, of a mountain having a pressure of 25 inches reigning at its top and a pressure of 30 inches at its base. The diminution in pressure which will occur during the ascent will be gradual, and the effect which will be produced on the aneroid by the time it arrives at the summit will be much less than the effect which would have been produced upon it if it had sustained a constant diminution in pressure of 5 inches for five consecutive hours. The loss upon the mercurial barometer will nevertheless be well marked, and at the end of the fifth hour, as the aneroid will then experience a constant reduction in pressure of 5 inches, it will lose rapidly. If a reading is taken on the summit immediately upon arrival there, the altitude ob- served by inspection of the aneroid will be somewhat (but not much) greater than the truth, 1 and it will be less than the altitude which will be observed if another reading is taken, an hour later. The result in 1 It is of course assumed that the aneroid has been correctly graduated, to read instantaneously against the mercurial barometer ; and it is also assumed that there is no change in atmospheric pressure, within the period involved, either at the higher or lower station. G 42 UPON THE USE OF THE ANEROID BAROMETER each case is due to the tendency to lose upon the mercurial barometer, which will still be actively at work at the end of the sixth hour, and at the end of that time let us suppose the descent is commenced. Immediately the descent is commenced pressure begins to be restored. The tendency to lose still continues, though with constantly diminishing force. For it is less potent at a pressure of 26 inches than it was upon the summit at a pressure of 25 inches, less potent still at 27 inches, and it dies out on return to normal pressure. But the moment pressure begins to be restored (as it does directly the descent commences) the tendency to recover sets in. This is feeble at first (when the restoration in pressure is slight) and it becomes stronger and stronger as normal pressure is approached. Thus, during the descent, there are two tendencies at work the tendency to lose in consequence of diminution in pressure, and to recover in consequence of restoration of pressure. These counter- act each other, and the effect is the recovery is retarded. I have found experimentally that, through the tendency to recover, a greater difference of pressure (or greater altitude) than the truth would always be obtained from descending observations (following ascending ones) if this were not checked by the tendency to lose ; and I have also found experimentally that retardation of the recovery causes descending readings to indicate less difference of pressure than ascending ones. 1 54. In Case 2, let us suppose the same mountain is ascended again, at the same rate, and that the traveller remains an entire day on the summit. In this case, if the traveller reads his aneroid immediately upon arrival at the summit he will, as before, obtain an altitude which will be somewhat greater than the truth ; and should he read it again at the expiration of twenty-four hours he will obtain an altitude consider- ably higher than he did from his second reading upon his first ascent, inasmuch as the aneroid will have been subjected continuously to a diminution in pressure of 5 inches for an entire day (instead of a single hour), and will have been losing upon the mercurial barometer during the whole time. At the end of this time, the tendency to lose will have become much less active (see table on p. 27) ; and, should he now com- mence to descend, the tendency of the aneroid to recover will be less checked than before. In consequence of this, over precisely the same ground, even if he comes down at precisely the same rate as upon his first descent, his descending readings will yield a result differing from that obtained 011 his first descent, and will indicate a greater differ- ence of pressure, and greater difference of level between the summit and the base than upon the first occasion. 55. For Case 3, let us suppose the same mountain is ascended again, at the same rate as before, and that the traveller remains an entire month on 1 Not, however, necessarily less than the truth. See later. Sometimes the opposition of these tendencies produces a perfect correction ; and the aneroid, at one point of its scale, will indicate the same differences as the mercurial barometer. See the Kew certificate given on p. 51. IN DETERMINATION OF ALTITUDES. 43 the summit. At the end of this time the tendency to lose will have ceased (or will be exerted very feebly). Thus, when the descent is commenced, the tendency to recover will be unretarded ; and, in consequence of this, over precisely the same ground, and during the same lapse of time as before, the descending readings will yield a result which will differ from those obtained upon the first two descents, and will indicate a greater difference of pressure, and greater difference of level than upon the first two occasions. 1 56. The remarks which have been made upon pp. 41-2 will probably be better apprehended after the tables have been examined that accompany the following experiments. These two experiments are designed to show that, in measuring precisely the same amount of difference of pressure upon several occasions, aneroids will yield dissimilar results upon each occasion, if the time conditions are dissimilar ; and also to exhibit the manner in which exact knowledge of the behaviour of aneroids may be turned to account in the measurement of pressures, and through them of altitudes. 57. EXAMPLE 1. The case is supposed of a traveller, provided with aneroids, starting from the level of the sea to ascend a mountain the height of which is unknown. It is supposed that no change in atmospheric pressure occurs during the expedition, and that the traveller proposes to deduce the altitude of his mountain from means of ascending and descend- ing observations of the differences of pressure indicated by his aneroids. He is supposed to occupy four hours upon the ascent ; to remain one hour upon the top ; and to descend from the summit to his starting-point (at the level of the sea) in two hours more. To read his aneroids just before his departure from the level of the sea ; upon the summit, one hour after his arrival there ; at the level of the sea immediately upon his return, and again at the level of the sea twenty-two hours later. What error, expressed in barometric inches, is he likely to fall into from his employment of aneroids in this manner ? For this experiment I used twelve aneroids, all of the watch size. 2 The aneroids were read ; pressure was then reduced gradually for four successive hours, at the rate 1-146 inches per hour ; it was maintained at 25*061 inches for one hour, and the aneroids were again read at the expiration of this (the fifth) hour. Pressure was then gradually restored during two hours. By the end of that time the aneroids arrived again at the pressure from which they 1 It is scarcely necessary to say that should an aneroid be used for ascending observations, immediately (or shortly) after it has experienced augmentation of pressure, the tendency to recover will retard the tendency to lose in precisely the same manner as the tendency to lose retarded the tendency to recover in the instances imagined in 53-55. Thus, if a traveller ascends a mountain upon two successive days, using the same aneroids on each occasion, and starts upon his second ascent (say) ten hours after his return to camp from the first one, he will from his second set of ascending readings obtain a less difference of pressure, and a smaller difference of level, than he did upon his first ascent. It may be pointed out here that the embarrassment caused by the ' recovery ' has greatly protracted these investigations. In the case of repetitions, it has sometimes been necessary to put instruments aside for months. 2 Because this is the size most commonly used by travellers. 44 UPON THE USE OF THE ANEROID BAROMETER ^ II i? s S s S 3 asfi-s C 3 3 > ^i^g-f COCDCDCDCDCOCOr-ir- ii ii (CO 10 II CN CN CM (M CN CN 73 43^ I^i c o la i< HSl OOOOOOOOOOiOO 66666666666 III IN DETERMINATION OF ALTITUDES. 45 had started, namely, 29-645 inches. The instruments were then again read, and were again read twenty-two hours later. In Col. 1 of the annexed table, the readings are given which were taken at the assumed sea-level ; in Col. 2, the readings upon the supposed summit, one hour after arrival ; in Col. 4, the readings immediately upon return to the starting-point ; and, in Col. 6, the readings which were taken twenty -two hours later. In Col. 3 the differences are given between the readings in Cols. 1 and 2 ; in Col. 5, the differences between the readings in Cols. 2 and 4 ; and in Col. 7, the differences between the readings in Cols. 2 and 6. At the bottom of the table I give the " Means of the Ane- roids," and the differences of those means from the difference of pressure indicated by the mercurial barometer. For the present, I invite attention only to the means of the aneroid observations in Columns 3, 5, and 7. The first will be found to be 0'021 of an inch in excess of, the second to be 0*030 of an inch less than, and the third to be 0*012 of an inch in excess of the difference of pressure indicated by the mercurial barometer. 58. EXAMPLE 2. It is supposed that our traveller's observations were considered unsatisfactory ; that it was said his work was too hurriedly per- formed, and that he should have taken at least two readings upon the sum- mit, several hours apart, and should have used the means of these two readings. He is supposed to have made another expedition ; to have started on the first day at 10 a.m., and to have reached the summit at 5 p.m. ; to have remained there until 5 p.m. on the second day ; and then, as before, to have descended in two hours to the level of the sea. He is supposed to have read his aneroids just before departure from the level of the sea ; to have read them again at 1 1 a.m. and at 5 p.m. on the day which was passed on the summit ; and again immediately on return to the level of the sea. What superiority would these observations possess over the former ones ? In this experiment I used the same aneroids as before, and read them against the mercurial barometer at the hours mentioned above. During the seven hours which are supposed to have been occupied by the ascent, press- ure was reduced at the rate of 0' 65 5 of an inch per hour, and during the descent it was restored gradually. In Col. 1 of the table (Example 2) the readings are given which were taken at the assumed sea-level ; in Col. 2 the readings on the supposed sum- mit, eighteen hours after arrival ; in Col. 4 the readings on the supposed summit twenty-four hours after arrival ; and in Col. 6 the readings upon return to the starting-point. In Col. 3 the differences are given between the readings in Cols. 1 and 2 ; in Col. 5 the differences between Cols. 1 and 4; and in Col. 7 the differences between Cols. 4 and 6. I again ask attention first of all to be given to the means of the aneroid observations in Columns 3, 5, and 7. The first will be found to be O'lll, the second to be 0-134, and the third to be 0-073 of an inch in excess of the true differ- ence of pressure observed from the mercurial barometer. 46 UPON THE USE OF THE ANEROID BAROMETER a i-, > <0 %%* I g & OiOiOiOiOiOiOiOiCSOOOt) CDCDJ>-CDCDOOOOJ> > O5COCOCD 4*<4!4t<4i<4t<4jO -1H 000*500000000 6006600066 j8.g dddddddddd 03 a s-J 03 > I $ S-S SIS |3 fl "^ o r< -Mean 4-5885. do. descending do. 4-572,' Error of aneroid No. 30, + 0-0045 of an inch. Thus, the worst result which was yielded by any single aneroid was less than two per cent in error, and the best one (like the mean of the whole) differed only 0*0045 of an inch from the mercurial barometer. Twelve aneroids and no more were employed in this experiment, and none of these were specially selected for it. They came from the stock of Mr. Hicks, and the maker was unaware of the nature of the tests to which they would be put. At the commencement of the experiment, care was taken that they were in what I have ventured to term "a state of repose," and no such results could have been obtained if this had been neglected. The reader will by this time perceive that not only are there various ways in which aneroids can be used but that the precise manner in which they can be most advantageously employed will be determined by the necessities of the case, I have first set before him some of the principal H 50 UPON THE USE OF THE ANEROID BAROMETER causes of error, and have then endeavoured to point out some ways in which these can be guarded against. He has seen that precautions are desirable in some instances, and are indispensable in others. He has seen that aneroids are at present constructed to read instantaneously, or nearly so, against the mercurial barometer, and he has been shown that they are very materially affected by variations in pressure, that the extent to which they are affected depends upon the greater or less lapse of time, and that they are unequally affected by these causes. 62. Some may wish to measure a mound in a few minutes or to bound in balloon ten thousand feet in an hour, and others to spend a day upon the ascent of a mountain or to pass weeks, months, or years at great elevations in the interior of continents. It is impossible to dictate the right course of procedure for every case that may occur. The best results will be attained by those who make themselves acquainted with the rules that govern aneroids in general, and with the particular behaviour of the instruments which may be employed. I offer the following remarks as hints rather than as directions. 1. When only a few minutes elapse between readings, such errors as may occur will most likely be due to errors of graduation. 2. When readings are taken a few hours apart (for difference of level) it will be advantageous to employ the means of ascending and descending readings ( 52, 61). 3. When days elapse between readings taken to obtain difference of level, it will be necessary to know the amounts which will probably be lose or recovered at the pressures experienced, during the length of time concerned ; and desirable to be able to determine index-errors ( 47). 4. It seems to me indispensable for all those who aim at correct measurement of altitudes above the level of the sea by means of aneroids to be at all times able to determine their index-errors. No one, I think, will regret paying attention to the following recom- mendations : 1. Prefer open scales. 2. Prefer aneroids of large diameter to the watch size. 3. Prefer aneroids which have been made some length of time. 1 4. Avoid working aneroids up to the extreme inferior limit of their scales. 5. Treat an aneroid with the same care as a watch. 6. Trust more to the scale of inches than to the scale of feet. Something may be gained by paying attention to the foregoing hints and recommendations, but the two essential requirements for those who use aneroids in the field upon prolonged journeys are 1. knowledge of the amounts which will probably be lost and recovered by their instru- ments, and 2. ability to determine their index -errors at any time. 63. I do not venture to suggest what form of certificate would be best adapted for general use. The settlement of this may well be left 1 Old aneroids (that is, aneroids which have been made for a number of years) are generally more sober and regular in their conduct than young ones. This fact is, I believe, already well known to instrument makers. IN DETERMINATION OF ALTITUDES. CERTIFICATE OP EXAMINATION 51 ISSUED BY THE KEW OBSERVATORY, RICHMOND, SURREY. ANEROID BAROMETER No. 899. by CART, London. Compared with the Standard Barometer of the Kew Observatory (reduced to 32 JW-) with the following results. Pressure- XV /f Correction to Aneroid with a pressure diminishing. f* 00 '. 0-00 06 Correction to Aneroid with a pressure increasing. -A*. -t~0 3* -+0.30 +0 i. 30 i +0. * It is probable that this correction will have become reduced to that first obtained after the instrument has remained at the normal atmospheric pressure for a short time. Note. When the sign of the Correction is +, the quantity is to be added to the observed scale reading, and when - to be subtracted from it. KEW OBSERVATORY, October 1888. G. M. WHIPPLE, SUPERINTENDENT. 52 UPON THE USE OF THE ANEROID BAROMETER to those whose business it is to attend to such matters ; but, whatever it may be, it should, I think, very materially differ from the certificate given overleaf, which is one recently issued by Kew Observatory. It is convenient to refer to this certificate for illustration of several of the points which have been touched upon. The aneroid No. 899, Gary, has twice passed through my hands on the first occasion in 1889, seven months after it had been tested at Kew, and on the second in 1890, after a journey in the Caucasus upon which it had been employed. In 1889, when it first came to me, I tested it under the air-pump in the ordinary manner, reducing press- ure inch by inch, and reading it against the attached mercurial baro- meter at each successive inch. 1 The time occupied in reducing it from 30 to 15 inches amounted to twenty -five minutes. The results of my comparisons agreed closely, at every inch of the scale, with those obtained at Kew. At 29 inches I found the aneroid read 0-073 of an inch too high (the corresponding result at Kew was +0'050), and at 15 inches it read 0'045 too low (Kew result -0-050). As the Kew certificate only goes to the nearest half-tenth of an inch, their results and mine may be considered to agree, and to be satisfactory, as they show that No. 899, Gary, upon being tested independently upon two occasions, seven months apart, behaved almost alike on the two occasions, and that the prognostication contained in the passage at the foot of the Kew certificate was correct. 2 I then reduced No. 899, Gary, to a pressure of 16 inches, and kept it at that pressure for eleven days. At the end of a week this aneroid lost 0*625 of an inch, and at the end of the eleventh day 0-736 of an inch upon the mercurial barometer (the Kew result was 0*050). Thus, if the owner of this aneroid had remained with it at a pressure of 16 inches for a week, and had employed the correction stated upon the Kew certificate, he would probably have under-estimated the pressure to the extent of 0-575 of an inch, which would have introduced a very large error into computations of altitude. He would have employed the error due to a diminution of pressure for the space of about twenty-five minutes instead of that arising in a week, and from this, I think, it sufficiently appears that it is desirable to present certifi- cates of examination in a form different from that which is given upon p. 51. 64. The readings of No. 899, Gary, which were taken during the Kew examination can be recovered by applying the corrections stated upon the certificate, and are given below. In Col. 1 we have the readings of the Kew Standard Barometer, in Col. 2 the simultaneous readings of the aneroid as pressure was reduced inch by inch, and in Col. 3 the simultaneous readings of the aneroid as pressure was restored inch by inch. 1 The mercurial barometer readings, for comparison, were reduced to 32 Faht. 2 This passage (marked by an asterisk) is a recognition of " the recovery. " DETERMINATION OF ALTITUDES. 53 1. 2. 3. Readings of K. O. Standard Barometer. Readings of No. 899, Gary, with pressure diminishing. Readings of No. 899, Gary, with pressure increasing. inches. inches. inches. 30-000 30-000 29-800 29-000 29-050 28-750 28-000 28-050 27-800 27-000 27-000 26-750 26-000 26-000 25-700 25-000 25-050 24-700 24-000 24-050 23-700 23-000 23-000 22-650 22-000 22-000 21-650 21-000 21-000 20-650 20-000 19-950 19-650 19-000 18-950 18-700 18-000 17-950 17-750 17-000 16-950 16-800 16-000 15-950 15-850 15-000 14-950 14-900 It will be seen that at no single inch do the aneroid readings in the two columns agree. Every reading in Col. 3 is lower than the corresponding one in Col. 2. It is apparent that in the short space of time during which this aneroid was submitted to diminished pressure (a space of time which probably did not amount to one hour in all), it showed a distinct loss upon the mercurial barometer at every inch of its scale. The differences between the readings in the two columns at each inch represent the amounts lost at those parts of the scale during the brief time that was occupied in reducing pressure to 15 inches and restoring it again. Thus, at 21 inches, although there is no error in Col. 2 (the aneroid reading truly against the mercurial barometer) there is a minus error of 0-350 of an inch in Col. 3. This amount (0-350 of an inch) is the amount which was lost during the time pressure was reduced inch by inch from 21 to 15 inches, and subsequently restored inch by inch to 21 inches, 1 and the same explanation applies to every other inch of the scale. As it is demonstrable from the certificate itself that the aneroid lost upon the mercurial barometer at every inch of its scale when pressure was diminished, I am unable to see the propriety of applying as correc- tions the amounts which are stated on p. 51 (in the column headed "Correction to aneroid with a pressure diminishing"). For example, at 23, 22, and 21 inches it is directed that there are no corrections when 1 It is probable that this part of the examination did not occupy so much as thirty minutes. 54 UPON THE USE OF THE ANEROID BAROMETER pressure is diminishing, yet the certificate itself shows that the aneroid lost 0*350 of an inch at each of those inches during the time it was under trial. From this it will, I think, be apparent that it is desirable to make some radical change in the style of Certificates of Examination. 65. The last point to which I wish to draw attention is the change in index -error which occurred in this instrument. When No. 899, Gary, was examined at Kew in 1888 it had a plus error of 0*050 in. at 29 inches, when the testing was commenced (mere. bar. 29*000, aneroid 29*050). At the termination of the examination this was changed to a minus error of 0-250 in. (mere. bar. 29'000, aneroid 28*750). The aneroid recovered, and by the time it came into my hands in 1889 the error at 29 inches was + 0*073. After my testing was done, I retained the instrument for a few days to watch its recovery, and when it left me it had at 29 inches a minus error of 0*179 (mere. bar. 29*000, aneroid 28*821). It was continuing to recover, and, by the time it was first used in the field, I have no doubt it was reading closely against the mercurial barometer. But when this instru- ment came to me the second time (in 1890), after having experienced large variations in pressure, its index-error at 29 inches was 0*779 of an inch (mere. bar. 29*000, aneroid 28*221). I re-tested it under the air-pump, and found that the major part of the error was carried along the scale. At 16 inches its index-error was - 0*596 of an inch (mere. bar. 16*000, aneroid 15*404). If it had been again kept at a pressure of 16 inches for a week it would again have lost a large amount, and this time the loss, so to speak, would have been piled up on the top of the index-error ( - 0*596), and in the aggregate would have exceeded a barometric inch. Although the change in index-error was obvious on return to London, it would not have been so apparent to an observer in the field unless he had been possessed of the means of determining index-errors. He might, and probably would have continued to apply the corrections stated upon the Certificate of Examination. The Kew certificate of No. 899, Gary, declares the errors which were remarked at each inch of the scale of this aneroid in Oct. 1888, upon its being reduced, inch by inch, as rapidly as possible. So long as there was no index-error at 30 inches, the corrections stated in the certificate were proper ones to employ, provided the aneroid experienced exact repetition of the same treatment. But they are not the proper ones to employ if the time conditions are different, or if there is change in the initial index-error. In the absence of directions to the contrary, it may be assumed by persons into whose hands similar certificates come that the corrections stated in them are good under all conditions and for all time ; and I have therefore thought it desirable to point out the limited value of this form of certificate, and again to dwell upon the necessity of workers in the field being able to re-determine index-errors of aneroids at any time. 66. To determine the index -errors of aneroids in the field one must have recourse to the mercurial barometer a mountain barometer, on the Fortin principle, by preference, if it can be carried. Upon some rough or prolonged journeys it is well-nigh impossible to transport IN DETERMINATION OF ALTITUDES. 55 mountain barometers in safety. It is, however, possible on all journeys, of whatever nature, to carry graduated (unfilled) barometer tubes and a bottle of mercury. In its simplest form, a mercurial barometer can always be available ; and by reverting to the practice of earlier travellers in employing plain or graduated tubes and filling them on the spot 1 one may combine the accuracy of the mercurial barometer with the conveniences of the aneroid. During the progress of the experiments which are recorded in this paper, several travellers have made important journeys upon which they have relied on aneroid barometers for determination of altitude, and they have sometimes said that their aneroid observations agreed very well with their boiling-point observations. I find this difficult to understand, and I think that the difficulty will be shared by the travellers themselves if they will submit their aneroids to artificially-produced diminution of pressure for some length of time, and note the errors which they will exhibit upon the mercurial barometer. I could not speak with any confidence upon this matter if there had been an exception to the general rule ; but I repeat that through- out the whole of my experiments in the workshop there has been no occasion upon which there has been one, and that every aneroid, without exception, which has been submitted to diminished pressure for a month, a week, or even for a day, has first lost in a marked degree upon the mercurial barometer, and has recovered upon restoration of pressure. I invite those who are able to do so to perform such experiments as have been described in these pages, to convince themselves of the serious nature of the mistakes which must be fallen into by travellers and others if they are led to suppose that aneroids will exhibit the same errors upon the mercurial barometer at the end of a week or a month as they do in a few minutes or hours ; and also to satisfy themselves that comparisons of a traveller's aneroids against the mercurial barometer, at natural pressure, upon return to the level of the sea after prolonged journeys in elevated regions, have not the value which is at present assigned to them. My best thanks are due to the various gentlemen who have aided me by lending aneroids, especially to Mr. Louis Casella and to Mr. J. J. Hicks for the facilities that they have rendered to an enquiry which might have proved damaging to their interests. It is to be hoped they will have their reward. I shall have mine if the publication of this paper leads to the more scientific use of a valuable instrument. 1 To be done in camp, and aneroids to be used when on the march. Although this is a rough method, excellent results have been obtained by it in the past, and it is likely to afford a better standard for comparison than will be evolved by boiling thermometers. At the best, the boiling-water method is a circuitous manner of arriving at a result which can be obtained with greater facility and by a smaller consumption of time in another fashion. See Travels amongst the Great Andes of the Eqitator, Appendix B. 56 UPON THE USE OF THE ANEROID BAROMETER NOTE UPON THE GREENWICH TABLE OF "CORRESPONDING NUMBERS OF ELEVATION." The table to which I referred at p. 39 (note) is intended to be used for the graduation of Aneroids. It starts from zero (level of the sea) and gives 3 1 inches as the corresponding height of the Aneroid or corrected Barometer (that is the Mercurial Barometer reduced to 32 Faht). This table is given (condensed and extended) upon the opposite page. Many instrument-makers have followed it literally, and made 31 inches their zero, or level of the sea. At the foot of this Greenwich Table the instructions are given which are reproduced upon p. 57 ; and, if they are followed, there is no objection to 31 or even 32 inches being made zero. But in practice many persons find it impossible to follow these instructions ; many others are unacquainted with them ; * and yet more would not follow them if they were acquainted with them. A considerable proportion of those who employ aneroids wish to obtain a fair notion of their elevation above the level of the sea by mere inspection of the height indicated by the index-needle, without reference to "lower stations," or being obliged to make computations of any sort. I have frequently pointed out to instrument -makers, and think it may be useful to say here, that these persons are unnecessarily led into error through the zero being placed at 31 inches. Inasmuch as the mean Barometer at the level of the sea, in all parts of the world, is much nearer to 30 than to 31 inches, 2 the adoption of 30 inches as zero would be more appropriate. Should this be done, the " corresponding numbers " will be those given in the following table (the mean of atmospheric temperature being 50 Faht.). Aneroid. Inches. Height in feet. Aneroid. Inches. Height in feet. Aneroid. Inches. Height in feet. 31-000 - 894 24-500 + 5520 18-000 + 13,923 30-500 - 450 24-000 + 6082 17-500 + 14,691 30-000 23-500 + 6656 17-000 + 15,481 29-500 + 458 23-000 + 7242 16-500 + 16,295 29-000 + 924 22-500 + 7841 16-000 + 17,133 28-500 + 1398 22-000 + 8454 15-500 + 17,998 28-000 + 1881 21-500 + 9081 15-000 + 18,892 27-500 + 2372 21-000 + 9722 14-500 + 19,817 27-000 + 2872 20-500 + 10,378 14-000 + 20,773 26-500 + 3381 20-000 + 11,051 13-500 + 21,762 26-000 + 3900 19-500 + 11,741 13-000 + 22,786 25-500 + 4429 19-000 + 12,449 12-500 + 23,848 25-000 + 4969 18-500 + 13,176 12-000 + 24,952 1 As they are not issued with the aneroids. 2 See Smithsonian Miscellaneous Collections, 153 ; entitled Tables Meteorological and Physical, section iv (hypsometrical tables), tab. xii, p. 85. IN DETERMINATION OF ALTITUDES 57 CORRESPONDING NUMBERS OF ELEVATION IN ENGLISH FEET, AND OF READINGS OF ANEROID IN ENGLISH INCHES (THE MEAN OF ATMOSPHERIC TEMPERATURES BEING 50 FAHRENHEIT). Height in feet. Aneroid. Inches. Height in feet. Aneroid. Inches. Height in feet. Aneroid. Inches. Height in feet. Aneroid. Inches. 31-000 6000 24-875 12,000 19-959 18,000 16-016 200 30-773 6200 24-693 12,200 19-813 18,200 15-899 400 30-548 6400 24-512 12,400 19-669 18,400 15-783 600 30-325 6600 24-333 12,600 19-525 18,600 15-668 800 30-103 6800 24-155 12,800 19-382 18,800 15-553 1000 29-883 7000 23-979 13,000 19-240 19,000 15-439 1200 29-665 7200 23-803 13,200 19-100 19,200 15-326 1400 29-448 7400 23-629 13,400 18-961 19,400 15-214 1600 29-233 7600 23-457 13,600 18-882 19,600 15-103 1800 29-019 7800 23-285 13,800 18-684 19,800 14-993 2000 28-807 8000 23-115 14,000 18-548 20,000 14-883 2200 28-596 8200 22-946 14,200 18-412 20,200 14-774 2400 28-387 8400 22-778 14,400 18-277 20,400 14-666 2600 28-180 8600 22-611 14,600 18-143 20,600 14-559 2800 27-973 8800 22-446 14,800 18-011 20,800 14-453 3000 27-769 9000 22-282 15,000 17-880 21,000 14-347 3200 27-566 9200 22-119 15,200 17-749 21,200 14-242 3400 27-364 9400 21-957 15,400 17-619 21,400 14-138 3600 27-164 9600 21-797 15,600 17-490 21,600 14-035 3800 26-966 9800 21-638 15,800 17-362 21,800 13-932 4000 26-769 10,000 21-479 16,000 17-235 22,000 13-830 4200 26-573 10,200 21-322 16,200 17-109 22,200 13-729 4400 26-379 10,400 21-166 16,400 16-984 22,400 13-629 4600 26-186 10,600 21-012 16,600 16-860 22,600 13-529 4800 25-994 10,800 20-858 16,800 16-737 22,800 13-430 5000 25-804 11,000 20-706 17,000 16-615 23,000 13-332 5200 25-616 11,200 20-554 17,200 16-493 23,200 13-234 5400 25-428 11,400 20-404 17,400 16-372 23,400 13-137 5600 25-242 11,600 20-255 17,600 16-253 23,600 13-041 5800 25-058 11,800 20-107 17,800 16-134 23,800 12-946 This Table is intended more particularly for the graduation of Aneroids with a circle of Measures in Feet concentric with the ordinary circle of Barometric Height measured in Inches. The circle of Feet is to be read off, at the upper and lower stations, by the Index ; and the rule for measuring the height will be : Subtract the reading at the lower station from the reading at the upper station ; the difference is the height in Feet. The Table supposes the mean temperature of the atmosphere to be 50 Fahrenheit. For other temperatures the following correction must be applied : Add together the temperatures at the upper and lower station. If this sum, in degrees of Fahrenheit, is greater than 100, increase the height by yinnr P ar t for every degree of the excess above 100 ; if the sum is less than 100, diminish the height by TT nnr part for every degree of the defect from 100. I PART 4.-RECAPITULATION. THE LOSS. All aneroids lose upon the mercurial barometer when submitted to diminished pressure. 1-11, 21-24, 26, 28-30, 40, 42, 51, 57-9. When diminished pressure is maintained continuously, the loss com- monly continues to augment during several weeks, and sometimes grows to a very important amount. 5, 11, 2124. The most important part of any loss that will occur will take place in the first week. 21, 25, 26. The loss which takes place in the first week is greater than in any subsequent one. 21, 23. A considerable part of the loss which takes place in the first week occurs in the first day. 28, 29. The loss may be traced in a single hour, and in successive hours upon aneroids with expanded scales. 30. The amount of the loss which occurs is different in different instru- ments. Some lose twice, or even three times as much as others. See tables upon pp. 17, 21-24, 26. The amount of the loss which occurs in any aneroid depends (1) upon the duration of time it may experience diminished pressure, and (2) upon the extent of the reduction in pressure. 31. The loss that occurs at a pressure which may be well within the range of an aneroid gives only an imperfect clue to the loss which may occur at lower pressures. It is commonly the case for aneroids to lose very largely indeed at the inferior limit of their range. 31 (note). THE RECOVERY. When pressure is restored, all aneroids recover a portion of the loss which has previously occurred. 16-18, 32-35, 37, 41, 51, 53-5, 60. Some aneroids, in course of recovery, gain more than they have previously lost. 35, 36, 37 (table). Minus index -errors are some- times lessened, 33 (table) ; plus index -errors are sometimes increased, 32, 33 (table), 34 (table) ; and minus index -errors are sometimes con- verted into plus ones, 33 (table). RECAPITULATION. 59 The recovery is gradual, and commonly extends over a greater length of time than the period during which diminished pressure has been experienced. 16, 32, 37. In aneroids which have been kept at diminished pressures for a con- siderable space of time (a week and upwards), the most important part of the amount that will be recovered upon their experiencing restoration of pressure will be regained in the first week. 33 (table), 35. The greater part of the recovery of the first week is usually accom- plished in the first day. 34 (table), 35. The recovery in the first hour is almost always larger than that in any subsequent hour. 35. TESTING. The errors which will probably be exhibited by aneroids during natural variations of pressure may be learned approximately by submit- ting them to artificially - produced variations of pressure, 19 ; but the one-hour test which has heretofore been commonly applied for ' verifica- tion ' is of little value except for determining errors of graduation, or the errors which will be exhibited at similar pressures in a similar length of time. 20, 49, 65. It is desirable that there should be some change in the present form of Certificates of Examination, 63-5 ; and that the change should be a radical one, 64. ON THE ANEROID AT FIXED STATIONS. Under certain conditions, an aneroid will not follow natural diurnal or hourly variations in atmospheric pressure with reasonable accuracy. 40, 41. The index of an aneroid may indicate diminishing pressure at a time when pressure is actually increasing. 40. ON SIMULTANEOUS COMPARISONS OF ANEROIDS. Aneroids left at upper or lower stations (with the view of applying corrections to simultaneous observations at lower or upper stations) cannot indicate diurnal or hourly variations in pressure even with approximate accuracy unless they have remained at the level of such upper or lower stations for several days. 42. ON ALTERATION OF INDEX-ERRORS. The index-errors of aneroids are seldom or never constant. 49. 60 RECAPITULATION. In consequence of the difference between the amounts lost and the amounts recovered, as well as through other causes, the index-errors of aneroids are liable to alter considerably. 16, 17, 3539. A large percentage of aneroids gain under restoration of pressure a greater amount than they have previously lost under diminution of press- ure ; and a large percentage recover less than they have lost. It is exceptional to find the loss exactly balanced by the recovery. 36, 37. In some aneroids there is a continual tendency towards the plus, and in others towards the minus direction. 38. Index-errors generally pass along the scale when aneroids are used rapidly. 43-45, 65. Index-errors of aneroids need to be re-determined from time to time. 47-49, 62, 65. How index-errors may be determined. 66. ON MEASUREMENT OF ALTITUDES. It is probable that large reductions will have to be made in the height of many positions which have been determined by aneroid. 27. Reasons for this. 21-26, 28-35, 40. Ascending observations of aneroid, as a general rule 1 (provided the instruments are in "a state of repose" when starting from the lower station), indicate a greater diminution of pressure, or a greater altitude than the truth. 3-9, 14, 15, 21-31, 40, 52, 57 (table), 58 (table), 59, 63 (certificate). The error (due to loss upon the mercurial barometer) in the indicated decrease of pressure is greatest when diminished pressure (increase of elevation) has been experienced continuously for about a month, pp. 12, 13, 17 (tables); 11, 21-24. In altitudes deduced from ascending observa- tions of aneroids, errors are likely to be greatest when the instruments have been subjected continuously to diminished pressure for a month and upwards. 21-26, 48. Under certain conditions, ascending observations of aneroid may indi- cate less diminution of pressure, or less altitude than the truth. 55 (note) Provided the instruments are in "a state of repose" when starting from the upper station, descending observations of aneroid, as a general rule, indicate a greater increase of pressure, or a greater difference of level, than the truth. 12, 55. The error in the indicated increase of pressure is greatest when 1 There may be exceptions to the general rule, arising from defects (such as bad graduation), or accidents which are not discussed here. RECAPITULATION. 61 diminished pressure has been previously experienced for a month and upwards. 55. Under certain conditions, descending observations of aneroid may indi- cate less increase of pressure, or less difference of level, than the truth. 57 (see table, Col. 5) ; 60, 64 (see Col. 3 of table). Descending observations of aneroid seldom or never agree with ascend- ing ones. 50. The reason of this. 53-60. To determine differences of level, observe the differences of pressure indicated on the scale of inches rather than the differences in altitude indicated on the scale of feet. 62. Means of ascending and descending observations are recommended in certain cases. 61, 62. Comparisons of travellers' aneroids against the mercurial barometer, at natural pressure, upon return to the level of the sea after prolonged journeys in elevated regions, have not the value which is at present assigned to them. 17, p. 55. THE END. Printed by R. & R. CLARK, Edinburgh. VOLCANIC DUSTS, FROM THE GREAT ANDES OF THE EQUATOR. 1. VOLCANIC DUST FROM COTOPAXI, Eruption of June 26, 1877, which fell at Quito after an aerial voyage of 34 miles. See Travels amongst tlie Great Andes, chapter vi. [" In the brown glass chips vactioles are abundant, many of them range from about *001 to '002 inch in diameter, but some are still smaller." Prof. T. G. Bonney, Proc. Royal Soc., June, 1884.] 2. VOLCANIC DUST FROM COTOPAXI, Eruption of July 3, 1880, which was ejected to a height of 40,000 feet above the level of the sea, and fell on CHIMBORAZO after an aerial voyage of 64 miles. See Trends amongst the Great Andes, chapter xviii. The finest particles Weigh less than one twenty-five thousandth part of a grain. 3. VOLCANIC DUST FROM THE TERMINAL SLOPE OF COTOPAXI, 19,500 feet above the level of the sea. This is the matter which is ejected daily by the Volcano. See Travels amongst the Great Andes, chapter vii. 4. VOLCANIC DUST FROM 15,300 FEET ON COTOPAXI, about 2000 particles to a grain. See Travels amongst the Great Andes, chapter vii. ["The granules com- monly range from '01 to '015 inch in diameter. The most abundant are minute lapilli of scoriaceous aspect ; in less numbers are glassy whitish and reddish granules with these occur fragments of felspar, augite and hypersthene." Prof. T. G. Bonney, Proc. Royal Soc., June, 1884.] 5. LAPILLI FROM 15,000 FEET ON COTOPAXI. See Travels amongst the Great Andes, chapter vii. [" Pumiceous lapilli ; pulverized glass ; and mineral fragments." Prof. T. G. Bonney, Proc. Royal Soc., June, 1884.], 6. LAPILLI FROM AMBATO (8600 FEET). Mainly colourless, vesicular pumice. Many of the fragments have entangled within them small microliths, and also plates of a pale greenish mica. The town of Ambato is built upon this deposit. See Travels amongst the Great Andes, chap. iv. . 7. FINE VOLCANIC DUST FROM MACHACH1 (9800 FEET), existing as a con- tinuous stratum ten inches thick, the product of some unrecorded eruption of, great intensity. Consists largely of felspar and hornblende. The finest particles are felspar and pumice. "Almost as soft to the touch as cottonwool." See Travels amongst the Great Andes, chap. v. 8. FINE PUMICE DUST FROM MACHACHI (9800 FEET). This forms beds many feet in thickness, and consists mainly of clear, colourless, vesicular pumice, which includes greenish mica, some in minute hexagonal plates. See Travels amongst the Great Andes, chap. v. These Volcanic Dusts form most interesting and instructive objects for the Microscope. Sold in bottles, price Is. 6d. each. Each sample contains one grain and upwards. Mr. Gregory can also supply fine and coarse dusts from ETNA, Eruption of 1879, collected by the late Professor Silvestri, of Catania ; a few specimens of TYPICAL LAVAS FROM COTOPAXI AND CHIMBORAZO including the HIGHEST ROCK OF CHIMBORAZO, obtained at 19,300 feet; and also PUMICE FROM PICHINCHA. Volcanic Dust from an eruption in St. Lucia in 1812, which fell on board ship oflf Barbadoes, price Is. 6d. Also pieces of drifted Pumice from Krakatoa, washed ashore in Madagascar, price Is. each. Any of the Volcanic Dusts sent post free by return on receipt of Post order for Is. 7d., or the set of eight Andean specimens, in box, post free for 10s. J. R. GREGORY, 88 CHARLOTTE STREET, FITZROY SQUARE, LONDON. UNIVERSITY OF CALIFORNIA LIBRARY This book is DUE on the last date stamped below. FinELSchedule: 25 cents on first day overdue 50 cents on fourth day overdue me dollar on seventh day overdue. NOV16 1947 REC'D LD MAY 1 1 1959 LD 21-100m-12,'46(A2012si6)4120 \TOR, 5.8. .8. O. SALYIX, F.R.S. DAVID SHARP, M.B., F.R.S. T. R. R. STEBBING, M.A. And a PREFACE by EDWARD WHYMPER. HOW TO USE THE ANEROID BAROMETER. BY EDWARD WHYMPER. I. COMPARISONS IN THE FIELD. II. EXPERIMENTS IN THE WORKSHOP. III. UPON THE USE OF THE ANEROID BAROMETER IN DETERMINATION OF ALTITUDES. IV. RECAPITULATION. With numerous Tables. To range with "TRAVELS AMONGST THE GREAT ANDES." The above three Works are sold separately. CHARLES SCRIBNER'S SONS, 743, 745 BROADWAY, NEW YORK. VJA 7 L AMOUNT I PAMPHLET BINDER Manufactured &y ilGAYLORD BROS. Inc. Syracuse, N. Y. Stockton, Calif. "hymper. &cr mra 'T:U a~ now to US bs.roiri6'ter 3 the aneroi ^ _ - JV294554; UttS