July 2 i, 1923J 



NATURE 



99 



The Transport of Rocks. 



May I ask Prof. Grenville A. J. Cole, through the 

 Tnedium of your cohimns, how far the authority for 

 the statement that " the Portuguese stone . . . was 

 brought in carracks round the Cape to build the 

 jutting fort on the coral shore of Mofambique " 

 (Nature, March 17, p. 353) is to be regarded as 

 trustworthy ? 



I first saw this fort in 191 1, and as recently as 

 September last year I walked all round it. I have 

 never been inside, but I am told by Portuguese 

 residents on the island that the same kind of stone 

 has been used throughout in the construction of the 

 fort. This stone is a sandy coral-rock, with occasional 

 small pebble bands. The country rock of Mo9am- 

 bique island is also a coral rock identical in composi- 

 tion and fossil contents — so far as one can judge by 

 hand specimens and very numerous exposures — with 

 that of which the fort is built. This material occurs 

 in vast quantities on the eastern coast of Africa, and 

 indeed on many tropical coast belts : it is well seen 

 at Mombasa and Zanzibar, which island, like that 

 of Mo9ambique, consists of little else. The coral-rock 

 is not the best material for constructional purposes, as 

 an examination of the external walls of the ^ 

 fort is sufficient to show. Can it be that this 

 material was shipped all round the Cape ? 

 It may be so, but I find it difficult to believe. 

 E. J. Wayland. 



Fort Portal, Uganda, May 3. 



at the surface of the lower electrode. This fact 

 points to the existence of a considerable gradient of 

 potential in this layer. 



In the course of an investigation of the radiation 

 in the spark I have found, by means of direct electric 

 measurements, the existence of a considerable 

 gradient of potential in the thin layer that surrounds 

 the electrodes when the sparking discharge takes place. 



The discharge of 952 sparks per sec, yielding an 

 effective current of 24 milliamperes, shows that the 

 change of the difference of potential depends upon the 

 length of sparks, as the accompanying diagram (Fig. i) 

 shows. If the sparks are so short that the thin 

 layers in the proximity of the electrodes, which yield 

 a metallic spectrum, are not yet divided, then there 

 exists a great gradient of potential (Fig. i, I). The 

 size of gradient depends first upon the nature of the 

 metal forming the electrodes. This is shown by the 

 two curves on the diagram for electrodes of platinum 

 and aluminium. 



At longer sparks, while among the above-mentioned 

 layers only a spectrum of gas appears, the gradient of 

 potential is much less (Fig. i, II) ; this does not 

 depend upon the nature of the electrodes. 



The intermediate space, marked by interrupted 



In reply to the interesting letter from Mr. 

 Wayland of the Geological Department of 

 Uganda, I beg to say that my authority for 

 the statement that the fort of Mozambique 

 was built of stone brought from Portugal is 

 the uninitialled article in the " Encyclopaedia 

 Britannica," nth ed., vol. 18, p. 949, where 

 we read : " There are three forts, of which 

 the principal, St. Sebastian, at the northern 

 extremity of the island, was built in 15 Jo 

 entirely of stone brought from Portugal." 



I have examined the coral-rock here and at 

 Mombasa, and, as Mr. Wayland states, it is 

 not attractive for building purposes. I cannot speak 

 as to the outer wall of the fort, and it may have been 

 rebuilt or refaced since 15 10. It would be interesting 

 now to pursue the matter in some detailed history of 

 Mozambique. Grenville A. J. Cole. 



Kv 



■K-A 



On Auroral Observations. 



It has been found that the green auroral line is 

 regularly visible in the clear night sky, and Lord 

 Rayleigh has discovered the remarkable fact that it 

 is more intense at Terling than in the north of England. 

 A cognate investigation, which, so far as I know, has 

 not yet been made, may be suggested to auroral 

 observers, namely, to examine how the intensity 

 changes at any one place throughout the night. The 

 observation is doubtless a difficult one, but might be 

 made by exposing a series of plates at different hours 

 on a succession of clear nights. It would be of great 

 interest to know whether or not the intensity remains 

 nearly uniform throughout the night hours. 



S. Chapman. 



The University, Manchester, 

 July 4- 



Gradient of Potential near Electrodes. 



In Nature of March 31, p. 431, Messrs. H. Nagaoka 

 and Y. Sugiura describe a method of observing the 

 Stark effect in the iron arc ; namely, in the thin layer 



NO. 2803, VOL. I I 2] 



lines, is difficult to examine. Sparks of both kinds 

 generally come into view. 



If V^ signifies the difference of potential for short 

 sparks, and Vj that for long sparks, we get for platinum 

 and aluminium electrodes about : 



/grad^VA =10 / grad. VA =5 

 Vgrad. V2/ Pt Vgrad. \\) PA 



The thickness x of the layer where there is a 

 considerable gradient of potential is small : 



Pt. Cu. Al. 



X -o.^ 0.4-0.5 0.7-0.9 mm. 

 This investigation is being continued. 



S. PlEls'KOWSKI. 



Physical Institute, 



University of Warsaw, 



Hoza 6g. 



The Tides. 



The notice in Nature of April 14, p. 508, of my 

 pamphlet on the tides implies that I have completely 

 misunderstood " the theory of the tide-generating 

 force on the principle of gravitation." Regarding 

 this I would like to present to your readers, very 

 briefly, just one point in that theory : Newton, 

 Herschel, and many other authorities compute the 

 principal tide-raising force as the difference between 

 the moon's attraction at the earth's centre and at 

 the earth's surface. Now this is exactly the method 



