Jan. 1 6, 1879] 



NATURE 



241 



cloud persistently hung over the island and prevented 

 anything like satisfactory observation, and had it not 

 been for Mrs. Gill's pluck, failure might have been the 

 result. It occurred both to her and her husband that the 

 cloud was only local, and to prove the truth of this sugges- 

 tion, Mrs. Gill undertook a journey along the coast, 

 starting about midnight, over great cinders and deep 

 rifts, to a point ahnost four miles south. The supposition 

 proved true, and amid many difficulties the observatory was 

 dismounted, and the instruments 'removed to the south- 

 west of the island, to a small inlet christened Mars Bay, 

 in memor}' of the expedition. But what a change from 

 "Commodore's Cottage," as the "Garrison" residence 

 was called. Ascension is an extinct volcano, and it is 

 now little more than a huge mound of cinders and dust. 

 On such floor did Mr. Gill pitch his tent, and on such 

 a base had he to erect his delicate instruments. The 

 discomforts attending his surroundings knocked him 

 completely up, but with the help of Mrs. Gill and the 

 doctor he was set on his feet again, and by the ministrj' 

 and companionship of the former the encampment was 

 made tolerable. Fortunately after all these hardships 

 and trials and doubts as to weather, the obser\-ations 

 at the critical time were completely successful, as were 

 a long series of subsequent comparisonj observations. 

 The captain of the island and his subordinate officers 

 deserve the greatest credit for the assistance and sup- 

 port which they gave to the enthusiastic astronomer and 

 his ever-helpful and cheerful wife. After the real work of 

 the expedition was completed Mr. and ]Mrs. Gill made 

 several excursions over the tiny island, and with the ex- 

 ception of an oasis on the summit of the " mountain," the 

 island seems dreary in the extreme, and Mrs. Gill failed 

 to find the neat square gardens and paved streets seen by 

 Sir Wyville Thomson. Altogether, on a very unpromising 

 subject, she has succeeded in writing a really interesting 

 and instructive book, telling us much about the islet and 

 its inhabitants, and still more about the circumstances 

 under which an important piece of scientific work was i 

 done. We strongly recommend it to the perusal of our ' 

 readers. ' 



i 



LETTERS TO THE EDITOR \ 



I 

 [ The Editor does not hold himself responsible for opinions expressed i 

 by his correspondents. Neither can he undertake to return, or \ 

 to correspond tuith the writers of, rejected nianuscnpts. No j 

 notice is taken op anonymous co7nniunications. 

 {The Editor urgently requests correspondents to keep their letters as | 

 short as possible. The pressure on his space is so great that it 

 is impossible otherwise to ensure the appearance even of com- 

 munications containing interesting and novel facts.'\ 



Schwendler's Testing Instructions for Telegraph Lines 



In the absence of my friend Mr. Schwendler in India, perhaps 

 I may be allowed to offer a few remarks on the notice of his book 

 on line-testing which appeared in Nature, vol. xix. p. 192. 



It must be remembered that the book is primarily intended for 

 the use of the officers of the Indian Tel^raph Department, and 

 that the conditions in that country differ very much from those 

 in England. Here the overland lines are in such positions that 

 any accident happening to them may be easily detected, but in 

 India the lines run in many parts through countries with few 

 inhabitants, and the distances between stations is sometimes very 

 great. Formerly when a breakdown or fault occurred, the line- 

 nders were sent out from the stations to find out what was the 

 matter, and Mr. Schwendler gave a very amusing account of two 

 of these native riding out from the two ends of the line to find 



a fault, meeting in the middle, salaaming and asking one another 

 if anything wTong had been seen. On receiving a negative reply 

 they salaamed and rode back, but the line was none the better 

 for it. It must be, to say the least, tedious to ride many miles 

 over a rough country staring at a wire on an Indian sky on the 

 chance of finding a dead snake across the wires or a bird's nest 

 on an insulator. These difficulties suggested the very systematic 

 line-testing now in vogue in India. 



It is unfortunate that the reviewer has shown such a disrespect 

 for mathematical formulae. There is no doubt that the book 

 swarms with them, but it is by means of these that Mr, 

 Schwendler has discovered many of the facts stated in the book, 

 some of which your contributor seems to doubt. I will, with 

 your permission, instance two cases. He writes, "Indeed it is 

 very doubtful w hether his proof that the sensibility of the bridge 

 method is greatest when the branch and the resistance are equal 

 is true. At any rate in our practice we find that the more dehcate 

 the galvanometer of the bridge the more sensitive and the more 

 accurate is our test." 



In the last sentence it is not quite clear what the reviewer 

 means by a "dehcate" galvanometer. I do not know that Mr. 

 Schwendler says that the galvanometer should not be delicate, 

 but he does say that its resistance should bear a certain relation 

 to the resistance measured. For a high resistance a galvanometer 

 of high resistance should be used, and for a low resistance one 

 of a proportionately low resistance. But the alteration of the 

 resistance should not be made by using only a portion of the 

 coils. This is fully explained on p. 22, where it is shown that 

 the coils should be connected either consecutively or parallel ia. 

 order to increase or diminish the resistance of the instrument. 

 Thus all the convolutions of the galvanometer are used, but as 

 the resistance is diminished in the second case a larger flow of 

 electricity takes place and a greater deflection is produced. The 

 results of these theoretical considerations are so readily tested by 

 experiment that it is surprising that the author of the notice 

 should have thrown doubt on their acctu^cy. I have therefore 

 thought that it might be useful to make some measurements 

 whici have fully confirmed the theory. I will not trouble you 

 with all the experimental niunbers of about 160 measurements, _ 

 which would be as "appalling" to the readers of Nature as 

 Mr. Schwendler's formulae are to the writer of the review, A 

 reflecting galvanometer with two coils was used. Connected 

 consecutively their resistance was 5590 Hnits and parallel 1405. 

 The results would have been more striking if the coils had been 

 of unequal resistance so that the parallel resistance would have 

 been less, or if the two halves of the bobbins could have been 

 connected parallel, which would have reduced the resistance to 

 about 700 units. The following table will show the results 

 obtained on measiu^ng three resistances with varying branch 

 resistances, and with the two arrangements of the galvanometer 

 the deflections of which were noted when a certain alteration of 

 the comparison coil was made : — 



k- These numbers show that the branches .should approximate to 

 the resistance measured and also that the galvanometer resistance 

 should be smaller when a small resistance is measured. Calcu- 

 lation shows that the most advantageous resistances of the gal- 

 vanometer in the three cases w culd be 2870, 8S0, and 95 respec- 

 tively. 



