Faraday to James Cosmo Melvill   9 June 1841¹

Royal Institution | 9th June 1841

Sir

I have received your letter of the 3rd instant, and also, from Mr. Daniell, the various papers it refers to; I have reperused Dr. O'Shaughnessy's² first papers³ and my report of the date of 5th September 1839⁴, and have now read Mr. Daniell's report of the 24th August 1839⁵ and Dr. O'Shaughnessy's further communications and remarks⁶. These, with the other papers, I have carefully considered and, as the general result, beg to say that I see no reason to alter even a word of my former report.

In saying so much, perhaps, I say every thing that the Honorable Court of Directors desire of me; for it is probably more the judgment of the individual, than the reasons for the judgment that is required. The latter, it is almost impossible for any man to give fully, in a case where there are arguments and reasons on both sides of a question; for the bearings of these become almost infinite in number, by reason of the variation in degree of force which they possess under varying circumstances; and at last it is the discretion, experience, tact, and caution of the person, which give to his opinion any worth, it may, in reality possess.

But as the present is a question which is referred back to me, I ought, perhaps, not to think it one in which the mere opinion without reasons will suffice; and yet I find a difficulty in going further into the matter:- for seeing and deeply regretting, a particular tone and character which some parts of the communications have acquired I am exceedingly anxious to keep myself quite clear from it, and may perhaps find this rather difficult. However, in what I may further say, I shall avoid referring to the opinions I agree with or differ from, as much as I can, yet without withholding my own opinion on any of the important points which are essential to the case and are intended by the court to be submitted to me.

There are, I think, three points under which all that is for consideration may be arranged.

1. Whether a good lightning conductor can cause a discharge where there would not be one otherwise

2. Whether when the electric fluid falls upon a conductor a part may not pass from it in the form of lateral discharge and occasion harm. 3. Whether at the moment the lightning conductor is struck, it may, by induction upon the gunpowder casks lined with copper or other metallic masses within the Magazine cause sparks to pass between them without any actua<l> lateral passage of lightning from the conductor<r>.

As to the first, I have already said that I think a good conductor, may, under certain circumstances, cause an electric discharge to take place where otherwise there might not be one:- that in fact it can virtually attract the lightning. It is to this quality that it owes much of its usefulness, if it have any; for unless it could divert the discharge (within certain limits) from a place where the lightning would otherwise have fallen, the conductor would have no use. But that it should cause a discharge upon itself and also at the same time upon a neighbouring object, as in the given case of Dr. Goodeves⁷ house, and that, not by a lateral discharge, is what I see no reason to expect either from principle, my experience, or such events as have come to my knowledge.

It is said that Dr. Goodeve saw the flash divide, part going to his own house in the Verandah of which he stood and part to the lightning conductor of the neighbouring house. When flashes occur so close to and over a person, taken at unawares as in this case, it is exceedingly difficult to distinguish exactly what takes place as to the direction of the electric flash, for there are many remarkable and confusing effects that simultaneously occur; still, admitting the division, I cannot but think the conductor acted as it ought to do rather than as it ought not, and that there would have been a much heavier explosion on Dr. Goodeve's house if the conductor had not in part averted it, though at the distance of 60 feet, than did occur under the existing circumstances.

The most important point however is 2. Whether when the electric fluid falls upon a conductor a part may not pass from it in the form of lateral discharge and thus cause harm. Before I make further remark upon this head I must repeat the words I used in the former report:- "I have no fear of lateral discharge from a well arranged conductor. As far as I understand lateral discharge it is always a discharge from the conductor itself. It might be very serious from a badly arranged conductor (and in fact makes them worse than nothing); but with a good lightning rod it can be but small and then not to badly conducting matter, as wood or stone, but only to neighbouring masses of good conducting matter as the metals; which, either, ought not to be there, or, if they are necessarily present, ought to be in metallic communication with the lightning conductor itself. I am not aware that lateral discharge can take place within a building when a lightning conductor outside is struck, except there be portions of metals as bell-wires, bolts &c, which may form an interrupted conducting train from the conductor to the interior."

That a lightning conductor well arranged as to its termination with the earth, especially if not of sufficient thickness, may give lateral discharge in the form of brushes and sparks, even when the quantity of electricity passing through it is not a thousandth part of that required for its fusion, or which the conductor could safely convey if alone, I can well believe and understand:- but for this to happen it requires an arrangement which I have already protested against above. The effect, when it does occur, is due to the resistance which, even, the best conductor makes to the passage of the electric fluid, and will be understood if I briefly describe a few experiments which I have made for my own satisfaction. We cannot always, it is true, say that our apparatus accurately represents natural circumstances, for the two sides of a Leyden jar do not correctly resemble the state of a cloud and the earth; the conducting power and aggregation of the charged surfaces are different, and so is the state of the medium between them, but the general principles are the same. In these experiments I used the mere spark of a large and good machine employing no Leyden jar[.]

A is the knob of a large prime conductor; B a metallic ball 6 inches in diameter, the distance from it to A being variable between the limits of 4 and 10 inches; C and D were wires of copper 40 feet in length, and each well connected at one of their ends with a large and extensive series of water pipes W; one of them was 1/20 of an inch in diameter and the other 1/10 of an inch[.]

When either of these was connected with the ball B and electric sparks passed from A to B, the electricity was conducted away perfectly by the wire; but when the end of the other D was brought to within a small distance of C in any part between E and B, a lateral spark passed from the conducting wire to the approximated wire, at the same time that the principal spark passed from A to B. The lateral spark was brighter and larger, according as the part of the conducting wire from which it was obtained was nearer to the ball B and further from the water pipe; and this is an important point, for the water pipe here represents the earth in a well arranged lightning conductor. When the wire D, was taken off from the water pipe and made fast to another set of pipes as its final discharge, still the same effects occurred. When a person standing on the carpet of the room tried to draw this lateral spark from C, by approximating the knuckle, he could scarcely obtain any traces, unless he held the wire D in his hand, and then the badly conducting matter of his hand could draw a very feeble spark. This lateral spark was still more beautifully shewn to depend upon the resistance of the matter of the wire, by taking a piece of wire O about 10 or 12 feet long and making it fast on to C at E and then bringing the other end near at some part towards B, as at H; for a lateral spark passed at H, shewing a certain division of the discharge through the two wires C and O. When a similar wire M was brought near to the wire C at K and at I, not being in contact at either place, then every time the lightning flash passed between A and B a small spark passed from the principal to the secondary wire at I and one from the latter back to the former at K. To complete this kind of observation the single wire C was removed, and replaced by a bundle of five parallel wires well connected at the two ends, and as their respective diameters were 1/10, 1/8, 1/8, 1/5, 1/5, of an inch, they were together equal to a copper rod above one third of an inch in diameter, and 40 feet long; representing, not inaptly, a lightning rod of small thickness. Still, when the flash passed between A and B, lateral sparks could be obtained from any or all of these wires at the parts towards the ball B.

I changed the discharger C, which represents the lightning conductor, and made it to consist of a fine copper wire only 1/42 of an inch in diameter. The same general effects were obtained; the lateral spark being, however, larger than before. I made the lightning distance from A to B vary from 8 inches to 2 inches and took the lateral spark continually from a certain spot near B, i.e. about 6 feet from it. The lateral spark was nearly half an inch long, and quite as long for the two inch lightning spark as for the 4 inch or 8 inch lightning spark; a result dependant on the circumstance, that, the 2 inch spark though containing less electricity is a quicker spark than the 8 inch or longer one:- time being here conceived in a manner, evident to an electrician[.]

I also made the lightning distance a constant interval of 3/4 of an inch, and then compared the effects of a small spark from the conductor only, with those of the denser spark of a large jar, to illustrate the influence of quantity. The lateral spark was the same length in both cases, but much brighter when the jar was used than when only the conductor spark was employed. In all these experiments with the fine wire C, an excellent discharger was used for D, (being a wire 1/5 of an inch in diameter) for the purpose of exalting the effects.

That these lateral discharges were really due to the cause I have assigned, namely the resistance in the metal to the passage of the electric fluid, and not to induction directly from the machine, was shewn by the following arrangement:- a part of the wire about I was completely sheltered from the effect of the machine by large uninsulated metallic plates, and yet the lateral spark could be obtained, though not so bright as when the plates were away. In the latter case inductive action & return discharge was combined with lateral discharge (pages 13, 14.)

Here then are lateral discharges and occasionally from a rod 40 feet in length, equal to 1/3 of an inch in diameter, well connected with a discharging system at the part representing the earth in a lightning rod; and that when the electricity sent through was certainly not a five thousandth part of that which the rod alone could have carried safely. What then (it may very properly be asked) can justify our placing such an instrument as this near a powder magazine?

In the first place it may be replied, that, all these striking instances of lateral discharge are obtained by using a good lateral conductor having, in all the cases but one, a good discharging termination with the earth and which therefore is itself a lightning conductor acting conjointly with the principal one. In the next place, where it was not in connexion with the earth or conductor (page 7) it was running parallel to it and, virtually, connecting parts at different distances from each other on the lightning rod, and so sharing its functions. In the next place it is found, as has been stated, that the lateral discharge is always greatest near the top of the lightning conductor, (or its representative,) and diminishes to nothing towards the lower end:- and in the last place, even with the finest wire and most exaggerated and best lateral conductor, the lateral spark contained but a small portion of the electricity of the principal or lightning spark.

Now it is by diminishing or removing entirely the influence of all these circumstances that a lightning conductor becomes a safe neighbour. It should consist of a sufficient body of excellent conducting matter, and it is considered that in this respect a copper rod one inch in diameter is enough:- it should be well connected by copper plates with the moist ground or water;- it should rise high above the building to be protected:- it should be placed near to it that what directive influence it has over the lightning may be used in protecting the building (page 14):- it should not come near masses of metal in the building, as a metal roof, or an iron column, or spout, or leaden pipe, or bell wires, or if it does, these should be metallically connected with it, and should themselves not go near to other metallic masses in the building and remain unconnected with them. In fact, the principal is, to make all that may act as conductors, and which would act if the lightning rod were not there, a part of the conductor; that no interruptions may occur in the electric course, and to give this course a free discharge into the ground:- the stored powder and such conducting masses being at the same time purposely separated from each other as far as they properly or conveniently can[.]

As a general illustration I may take Dr. Goodeve's house with its vertical window bolts, and say, that, if a good lightning conductor were fixed in that house and made to run parallel with the bolts at the distance of a foot or so, it is very possible, that, when lightning struck the conductor it might discharge in part to the bolts and much harm might be done; but if those bolts were in good metallic communication with the lightning rod no harm would be done:- and further, if the lightning rod were away, it is almost sure that the bolts without them would be struck, as in fact appears to have been really the case; for I cannot think the striking was due either to lateral discharge or to the attractive influence of the lightning rod standing 60 feet off.

A question has arisen with regard to the surface & the mass of a lightning rod: in the present case the question is of a mixed nature. As a conductor, surface has no influence over the power of the lightning rod; and copper plates or ribbons may not be substituted for a copper rod of equal superficies. The conducting power is as the square of the diameter in round rods, or, in fact, directly as the sectional area whatever the surface may be. A copper rod an inch in diameter is considered as sufficient, the conclusion being the result of general observation of natural and experimental phenomena. Flashes of lightning, though very awful and exceedingly intense in action, have not so much quantity of electricity in them, as many phenomena which continually pass unobserved before us; and frequently cannot fuze copper wires 1/5 of an inch in diameter or even such as are only 1/8 or 1/10 of an inch thick.

But as every metal opposes resistance to the passage of electricity, and is therefore an imperfect conductor, so surface, and the shape of that surface, has an influence and does to a certain extent affect the lateral discharge. In fact whilst the conductor, as our wire C, (page 6) can give a lateral spark, the electricity, which can cause that spark, is, in relation to external bodies, governed by the laws of statical electricity; and hence the influence of surface. Its effect however in a lightning conductor, unless purposely exalted and sought for can be but small.

The third point is:- whether at the moment the lightning conductor is struck it may, by induction upon the gunpowder casks lined with copper or other metallic masses within the magazine, cause sparks to pass between them without any actual lateral passage of lightning from the conductor? - Such a case would, according to my view, be a case of returning stroke, and be distinct from lateral discharge: the principles of such a case are easily illustrated. Suppose A a charged electrical surface, either, that of a cloud or of a prime conductor, B a lightning rod well arranged, and C a couple of metallic cases standing one on the other but separated by a sheet of paper, a cloth, or a thin piece of wood. As the surface A becomes charged positively, (we will assume it to be so) the upper end of the lightning rod B, and the upper case C will become negative by induction. Upon the discharge occurring to the lightning rod, the induction upon C would cease more or less, and it is possible that at that moment a spark should pass between the upper and lower metallic cases. But this can only happen if the cases C have been exposed to the induction of the cloud or machine as it rises in intensity, before the spark passes:- if they are protected from this as I think they are most perfectly by being within a building and under a roof then no case of this kind can occur[.]

If the metallic case or other mass of good conducting matter be both exposed to the inductive action of the charged surface, and also near the lightning rod, as at E then the effects of return and lateral discharge may be combined; for before the discharge on to the lightning rod, E will become negative by induction, and, at the moment of discharge a lateral spark will pass from the conductor to it, to restore its state (pages 8, 9):- a metal roof unconnected with the neighbouring conductor might present this case. These are results very easily obtained with electrical apparatus and machines.

I hasten to bring this letter or report to a conclusion. The whole question is whether such a low object as a powder magazine is safer with or without a lightning conductor. I cannot say that I have had very much experience of very powerful thunder storms, or such as may compare with those in India, but I know that in this country, very low objects are struck, and, therefore, I should protect by a lightning rod, such as I was very anxious to preserve. In place of increasing the number of those around a building I would rather increase the height of the one or two used, by a few feet. The distance from the magazine I would make about, but not further than, three feet[.] The other points mentioned in my former report I would again beg to urge your attention[.] The water tank at the Mazagon powder works would in my opinion afford sufficient discharge provided the communication of the lightning rod with it be made by metallic plates immersed in the water[.]

I have the honor to be | Sir | Your Obedient Humble Servant | M. Faraday

James C. Melvill Esq | Secretary | &c &c &c | East India House

Endorsed: Protecting effect of Conductor [word illegible] by Snow Harris - by Discharges upon the model of a mast, hollow & containing Detonating Powder - the wire was joined but no effect was produced on the powder.