Oct. 15, 1885] 
months, and in some cases even twelve months beforehand. The 
facts brought forward in that paper were of such a nature that, 
as will be readily understood, I wished very much they could be 
found to occur generally, But it was undoubtedly better to restrict 
their application to the area and period dealt with in the paper. 
It having been shown, however, that at one period and over a 
certain area quantitative relations had existed between previous 
and subsequent barometric variations, it is natural to suppose 
that quantitative relations may be found to exist at other periods 
and over other areas also. The question arises, Can the facts 
brought forward in the above-mentioned paper serve as a guide 
to future investigation? I think to a certain extent they can. 
The paper pointed out that there was a remarkable approach 
to an annual symmetry in the abnormal variations of the baro- 
meter in Western India during many of the years under observa- 
tion. It supposed that this symmetry would have occurred 
every year during that period had it not been masked by 
larger variations of another character; and it was mainly by 
acting on this supposition and noting the departure from 
symmetry in any given year, and by considering that departure 
as being an index of the variation that was about to come, that 
the position of the barometer in the subsequent year was calcu- 
lated. The paper attempted to explain the occurrence of this 
annual symmetry in two ways: (1) By supposing it to be a 
constant phenomenon connected with the annual double oscilla- 
tion known to be present in the normal barometric curve ; and 
(2) by supposing it to be a chance phenomenon, characterising a 
phase in the march of barometric variations, and persistent 
during the period dealt with, but not necessarily to be found in 
any other period. After further reflection I am inclined to 
believe that the latter is the correct explanation. 
And here I think may be a guide to future investigation. It 
seems very likely that barometric variations may always be 
passing through phases which are persistent for several years, 
And, during the continuance of each phase the abnormal baro- 
metric curve will necessarily approach more or less to a certain 
annual type. In the cases dealt with in my paper that type 
chanced to be of a symmetrical form, sufficiently remarkable to 
strike the eye at once. The regularity of its form made it com- 
paratively easy to be dealt with. An irregular type would of 
course be less easy to recognise and less easy to be dealt with. 
But it is obvious that if such types do exist and persist for several 
years in succession, then, by catching the type as the barometric 
phase comes in and by noting the departures from it each year, 
in a manner similar to that adopted with the symmetrical type I 
had to deal with, these departures may serve also in a similar 
manner as indices of the coming variations. Of course the 
methods of calculation would have to be purely arbitrary and 
specially devised for each barometric phase. If barometrical 
curves would yield to strictly mathematical methods, the problem 
of season-forecasting could be regarded as in a fair way of being 
solved. But it has never yet been found possible to resolve 
them entirely into regular periodical oscillations ; and I believe 
they will always have to be arbitrarily dealt with. 
Melbourne, July 21 A. N. PEARSON 
Transmission of Sound 
In connection with the subject of mechanical telephones, 
which has been occupying public attention lately, there is a note 
by Mr. Miller in a recent number of NaTuRE, regarding cer- 
tain experiments made in 1878 on the propagation of sound. 
With reference to this, Prof. Wernhold, of Chemnitz, writes to 
me, saying that as early as 1870 he had shown that human 
speech could be transmitted very distinctly through stretched 
wires or threads, and mentions that the results of his researches 
were published in an article on ‘‘ The Transmission of Human 
Speech through an Iron Wire,” in Carl’s ‘‘ Repertorium fiir 
Experimental Physik,” Band vi., Serie 168. As your corre- 
spondent will probably like to refer to this, may I ask you to 
kindly publish this letter ? W. E. AYRTON 
Central Institution, Exhibition Road, London, October 12 
Are there Rabbits in the Western Islands? 
PRoF. THOROLD ROGERS in his interesting book on ‘‘ Work 
and Wages” mentions the relatively high value of rabbits in the 
thirteenth century, and suggests that they were then a recent 
introduction to England. It is well known that several islands 
on the west coast of Scotland haye no rabbits upon them—for 
Nat RE 
57 
instance, Kerrera, which seems to point to the same conclusion. 
It would be interesting to know whether this is really the fact or 
not ? HERBERT ELLIS 
112, Regent Road, Leicester, October 4 
THE HELL-GATE EXPLOSION 
ROBABLY the largest chemical mechanical experi- 
ment ever thought of was successfully performed 
last week in New York Harbour by the removal of the 
obstruction known as Hell Gate, or Flood Rock, a con- 
siderable-sized island, as stated by the papers, about nine 
acres in extent, in Long Island Sound. The agent em- 
ployed for this immense engineering work is a preparation 
or preparations of nitro-glycerine, and there is no doubt 
that this is the only explosive compound which could 
have been used for the purpose on account of the very 
enormous quantity required and the peculiar nature of the 
explosion of this substance. All the compounds or prepara- 
tions of nitro-glycerine produce by explosion whatare known 
as local eftects only, as distinguished from gunpowder, 
the effects of which are much more gradually developed 
on ignition, but extend, owing to the slower and larger 
wave of disturbance, to a much greater distance. ‘The 
legitimate use of nitro-glycerine is for purposes such as 
this, where a disruptive action is required. 
The operations leading up to the final explosion have 
been some years in progress. They have consisted in 
forming a system of tunnels at a considerable depth 
under low-water level in the solid rock, and the charging 
of these tunnels with dynamite and mixtures known as 
rackarock, of nitro-glycerine with compressed gun-cotton. 
Twenty-four galleries were driven through this island, 
some of them 1209 feet long, and these were intersected 
by some forty-six others. These tunnels were about 10 
feet high and 8 feet wide, and the roof of rock above 
them varied from 10 to 25 feet in thickness. The quantity 
of rock to be removed by the explosive was about 275,000 
cubic yards, the quantity removed by tunnelling being 
about 80,000 cubic yards. A good deal of trouble has 
been occasioned during the course of the mining work 
by fissures, which have had to be stopped by 
wooden plugs in most instances. The explosive was 
charged into holes drilled into the roof and supporting, 
walls and pillars at different angles, with a view to dis- 
rupt the strata of rock as much as possible. 
The holes to be charged were about 9 feet in length and 
2% inches in diameter. The holes were charged first with 
the blasting gelatine or rackarock and filled to the ends 
with a dynamite cartridge, to which the detonator and 
electric wire were attached. In all fourteen thousand 
cartridges of a total weight of fourteen tons were em- 
ployed. Near observers describe the explosion as being 
accompanied by a dull roar, but with only the slightest 
shaking of the ground, even at a moderate distance. An- 
immense quantity of water was bodily raised up to heights 
estimated variously at 150 to 200 feet. 
The results, as far as can be ascertained, are very 
satisfactory, the rock having been very thoroughly 
broken up, so that it can easily be dredged away. 
After the example of an experiment on this scale, 
carried out without the least accident, perhaps it may 
occur to those in authority that we have on our own 
coasts dangerous rocks, not of the extent of Flood Rock, 
which might with immense advantage be similarly 
“ chemically” removed. 
Had gunpowder been the only explosive available, at 
least five times the quantity by weight of the nitro- 
glycerine preparations used in this experiment, would 
have been necessary and the results would not have been 
by any means so local or perhaps so satisfactory. 
After this the engineer may find it to his advantage to 
cultivate more the acquaintance of the chemist and his 
products than has been hitherto the case. 
