76 



APPENDIX, No. III. 



by more labour, or by more circuitous paths, 

 reached the same point to which his admirable 

 rules would at once have led us. 



From having carefully studied Captain Kater's 

 works before leaving England, we had conceived 

 ourselves sufficiently qualified to undertake a 

 course of experiments at once. In this, however, 

 we were mistaken ; and the consequence was, that 

 of two extensive series made at Valparaiso, neither 

 proved sufficiently accurate to deserve notice. 

 The experience, however, gained in the course of 

 these operations, enabled us ever afterwards to 

 proceed with confidence. And here it may be 

 well to suggest the advantage which, on future 

 occasions, might arise from having the whole ex- 

 periment performed in England, by the person 

 who is afterwards to repeat it abroad, not under 

 the hospitable roof of Mr. Browne, to whose in- 

 valuable assistance every one who has attended to 

 this subject is so deeply obliged, but in the fields, 

 and with no advantages save those which he could 

 carry with him. He would thus, in good time, 

 discover omissions in his apparatus, which are not 

 to be supplied abroad, and be aided in surmounting 

 difficulties before he had sailed beyond the reach 

 of appeal. 



The first series of experiments was made in 

 London. The next was made thirty-two miles and 

 a half north of the equator, at one of the Gala- 

 pagos, a group of islands in the Pacific, lying 

 upwards of two hundred leagues west from the 

 continent of South America. It was intended 

 that a station should have been chosen immediately 

 under the line, but the ship being swept to leeward 

 in the course of the night by a strong current, this 

 object could not be effected without losing more 

 time than circumstances admitted of being spent 

 in that quarter. 



The spot chosen for the experiments lies near 

 the extremity of a tongue of land running into the 

 sea at the south end of Abingdon Island, where 

 it forms the western side of a bay, about a mile 

 across. The point is a stream of lava, which, in 

 former ages, had flowed down the side of a peaked 

 mountain, standing in the middle of this end of the 

 island. The summit of this peak is between two 

 and three miles from the station, in a direction 

 nearly north, and is about two thousand feet high. 

 It slopes rapidly at first, so as to form a tolerably 

 steep cone, terminated by a broad and gently- 

 sloping base of a mile and a half. The sides of 

 the mountain are studded with craters, or mouths, 

 from whence, at different periods, streams of lava 

 have issued, and run down to the sea, where they 

 have formed sharp projecting points, such as that 

 on which we now fixed our station. The western 

 face of the island presents a cliff nearly perpen- 

 dicular, and not less than a thousand feet high ; it 

 exhibits a rude stratification of lava, tuffa, and 

 ashes, materials which characterise the fracture 

 of ancient volcanic mountains. I am thus minute 

 in describing this island, that the reader may be 

 enabled to judge how far its density may have 

 modified the results of the experiments. It is ten 

 or twelve miles long ; the north end being a con- 

 tinued system of lung, low, and very rugged streams 

 of lava ; the peak standing about one-third of the 

 whole length from the southern exti'emity, where 

 our station was. The rock, at different places 

 not far from the station, was found to be full of 



caves, into which the tide flowed through subter- 

 ranean channels ; the outer crust of the stream 

 having, as usual, served as a pipe to conduct the 

 lava off : it is therefore probable that our founda- 

 tion may not have been the solid rock : a circum- 

 stance which, taken along with the general hollow 

 nature of volcanic districts, and the deepness of 

 the surrounding ocean, renders these experiments 

 not so fit to be compared with those made in 

 England, as with others which may be made here- 

 after on a volcanic soil. 



The range in the temperature, in 24 hours, was 

 from 74° to 91° ; and, as we were obliged to place 

 the instruments in a tent, the thermometer rose 

 greatly in the day-time, and fell as much at night, 

 but unfortunately without much uniformity. On 

 the first day of observing coincidences, a set was 

 taken after breakfast, and another before dinner ; 

 but it was soon seen that this would confine the 

 observations exclusively to the hot period of the 

 day ; it was therefore determined to take in future 

 one set as soon after sun-rise as possible, in order 

 to have a result in which the performance of the 

 pendulum should be modified by the whole night's 

 continued low temperature ; and another set 

 towards the close of the day, to obtain a result 

 partaking in like manner of the influence which 

 the whole day's high temperature might have on 

 the length of the pendulum. We also endeavoured 

 so to arrange things, that we might catch a suffi- 

 ciently long period of uniform temperature during 

 the interval of observing, that all the coincidences 

 of each set might be taken with an unvarying 

 thermometer. By these arrangements it was 

 hoped, that although no one experiment could 

 produce strictly correct results, the errors of the 

 morning and evening observations, being of a con- 

 trary nature, might counterbalance one another ; 

 that the mean, in short, between observations 

 taken in the hot and in the cold periods of the day, 

 would probably give such a result as might fairly 

 stand by the side of rates deduced from transits of 

 stars, the intervals between observing which, in 

 like manner, included the very same extremes of 

 temperature. 



It should be carefully borne in mind, that the 

 real desideratum, as far as respects rate, is not to 

 know what is the aggregate loss or gain of the 

 clock in twenty -four hours ; or, in other words, 

 the mean rate ; but the actual rate at which the 

 clock is going during the particular period of ob- 

 serving : That is to say, the number of beats, and 

 parts of a beat, which, were the clock to go on 

 uniformly from that instant, would be indicated 

 by its dial-plate, in 24 hours of mean time. As 

 the method of transits of stars, however, gives only 

 the average rate, or that due to the middle point 

 of time between the transits, we sought, by the 

 arrangements above stated, to obtain, in like 

 manner, average results, by taking the mean of 

 observations with the pendulum made at the ex- 

 treme temperatures. 



One thermometer was suspended, so that its 

 bulb stood an inch in front of the middle part of 

 the pendulum, while another was hung lower 

 down, between the clock-case and the pendulum. 

 The average temperature at night was 74°, and in 

 the daytime, from 8G° to 90° ; the latter, as I have 

 said, depending principally on the state of the 

 sky. The allowance for expansion was made 



