CO885-(6-7) — Page 307

55

35

312

PUBLIC RECORD OFFICE

Reference :-

C.O. 885

6

ALLY WITHOUT PERMISSION OF THE BE REPRODUCED PHOTOGRAPHIC- COPYRIGHT PHOTOGRAPH-NOT TO

PUBLIC RECORD OFFICE, LONDON

17 November 1896.]

675. That

Mr. SIEMENS.

Mr. Murray-continued.

deteriorates? - I

showed you yesterday. That sample seems to have disappeared--a sample of the French cable which has been 16 years in the water (sample produced); you see that is in very good preser- vation, you see the outer-coverings are there, and you see that even the tarring-or rather they were galvanised wires that have got black, but at the same time it is in a perfect state of preserva-

tion.

676. Still if it does deteriorate you are not any better off by the fact that the core and the in- sulation improve ?-Well, I do not think I would absolutely agree with you, because you see you are telegraphing through the conductor, and as long as the conductor is insulated it does not matter whether the outer covering is good or bad; that does not influence the telegraphing, and the outer covering of steel wires is made so heavy and is made so strong that you can take up the cable and repair it, so that if the cable remains in that condition that it does not want repairs, it does not matter what becomes of the outer covering.

677. But sooner or later if the outer coverings went, the insulation would go?-I do not think so; not at such depths, and I am sorry- I might have brought it perhaps Mr. Preece can show you a sample of that core which was laid in 1851 by Mr. Brett across the Channel. The Post Office cable ship "Monarch" picked up some by accident, and I got a piece and you will find that the core gutta-peroha

looks rather better than those manufactured now-

a-days if anything. There is absolutely nothing to deteriorate in the gutta-percha if you can obtain the condition that there is no electrical fault. As I told you before, if there is the slightest beginning of a leak the cable will go sooner or later. If it is a small leak of course it will take a long time, but it is sure to go ; but if there is no leak, if you can make your cable in the first instance electrically perfect, then I do not know of any cause which should deteriorate it.

678. Then I should like to ask what is the object of the outside covering ?-The object is not to injure the core during the laying, and to have the possibility of repairing it.

679. It does not protect it while it is lying at the bottom then?-Oh, no. I take it that that is only for the purpose of laying it without injuring it, and of picking it up afterwards if necessary.

680. For all other purposes you might just as well have simply the core and the insula- tion?-For all other purposes.

Mr. Brett was perfectly right. He simply put down the core, which was all that was wanted for para- graphing; you only want the copper con- ductor and the insulation; you want nothing

else.

681. And both of these improve rather than deteriorate?-Well, the conductor, of course, is not touched, but the gutta-percha is certainly not deteriorated but rather improves.

Chairman.

[Continued.

682. I think, Mr. Siemens, you said that you proposed per nautical mile 500 lbs. weight of copper and 320 lbs. gutta-percha ?—Yes.

683. And that, you calculated, would give speed of 15 words per minute?—Yes.

a

684. Well, we had evilence yesterday of a proposal of 522 lbs. of copper and 368 lbs. gutta- percha, calculated to give a speed of 12 words per minute. How can you explain that discre- pancy ?-Ah, I am very glad I have brought the evidence to show you; that is, that we are of opinion that our cables are very much better, through reasons which I do not want to tell, than anybody else's, and if I may hand you this table- this is a comparison of speeds of the Atlantic cables. The usual way is to calculate the speed by the retardation co-efficient, being the product of capacity by conductor resistance. That gives you an indication of the speed the cable should give. Now, that retardation co-efficient is given in Column 8 of this table, and in Column 9 are given the calcu- lated speeds, taking the 1874 cable at 20 words per minute. Then, in the last column, are the actual speeds, and we have taken the actual speeds from official reports which are quoted underneath the table. For instance, with regard to the 1881 Gould cable, you will find that the co-efficient of retardation is 7-8, and the Anglo- American cable of 1874, which is taken as a standard, has a retardation co-efficient of 3:5; and the Anglo-American cable gives a speed of 202 words; consequently, accord- ing to calculation, the Jay Gould cable ought only to give 9.5 words, but as a matter of fact it gives 27. Then if we come to the last cables which I have filled in in ink you will see that the retardation co-efficient of the Mackay- Bennett is 46, and of the Anglo-American is 24. That is also taken from published figures, and the speeds of the cables should be 15·19 and 29; or the Anglo should be about double the speed of the third Commercial. Well, they have im- proved in their manufacture, and I could tell why, but I will not, so that instead of 29 words they get 47; therefore their speed is 50 per cent. better than it would be by calculation. But you see the Commercial cable, instead of 15 words, gives 42, that is nearly three times as good. This table also gives the lengths, and I see that I made a mistake yesterday. It is one of the Jay Gould cables which is the longest that is 2,563. Somebody asked yesterday. the longest cable at present existing.

Mr. Murray.

That is

685. What is the 1875 one?-The 1875 cable ie 2,423. If you take the Mackay-Bennett at 2,161 and the Anglo-American as only 1,850 miles long, if you calculate out what the Mackay- Bennett would give if it had only the length words; so with a core of 500 lbs. copper and of the other, you will find that it would give 50

320 lbs. of gutta-percha against a core of 650 lbs. of copper and 400 lbs. of gutta-percha it would give nearly ten per cent. more speed; but if I may do so I would strongly recommend to you to get Dr. Alexander Muirhead to talk to you on

17 November 1896.]

Mr. SIEMENS.

Mr. Murray-continued. that subject. He is the inventor of the duplex system, and he has applied this duplex to all the Atlantic cables, I think at least to nearly all, and therefore he has impartial information to tell you about this difference in speed between our cables and other cables.

Chairman.

686. Will you put in the table you have just been reading from, or a copy ?-If you will allow me, yes. (Table put in.)

687. And the specification ?—Yes. (Specifi- cation put in.)*

Mr. Murray.

688. Is there any definite relation that you can state between the weight of the insulation and the weight of the core?-Well, we have arrived at that also experimentally, and we have done it in this way we have counted the faults which occur in the cores while they were being manufactured, and we have found out that if the faults exceed a certain number per knot, the gutta-percha is rather thin, that we ought And then to put more gutta-percha on. we have found out if we have taken the conductor, say a conductor weighing 100 lbs., and covered it first with 70 lbs., then 80 lbs., then 90 lbs., then 100 lbs., and then 110 lbs., and SO

at last the number on,

same

of faults per knot, which occurred during the manufacture, remains constant, there- fore we say the minimum weight at which this constant number of faults occurs must be the minimum weight which we dare put on a con- ductor having that weight; and in that way we have made an experimental table of all the cores which we have ever manufactured, showing the number of faults which have occurred, and in that way we have experimentally fixed the minimum weight of gutta-percha which we should put on a given conductor. Generally speaking, we put a little more on to be on the safe side. For instance, we have found that with 500 lbs. of copper we could do with 300 lbs. of I mean gutta-percha, but we put 320 lbs. on; simply allowing a margin.

689. Because I see that in your shorter lengths the insulator exceeds the weight of the core in many cases?—Yes; that is because you see the specific weight of the copper is so much greater than that of the gutta-percha. The space occupied is not in proportion to the relative weights.

690. Then I take it that the carrying capacity of the cable cannot be deduced simply from the weight of the core and insulator and length, but you have to consider the quality of the cable besides ?-Yes, the manufacture.

we

691. The manufacture?-I mean we claim and as I should prefer if you go to independent testi- mony and ask Dr. Alexander Muirhead: claim that our cables are, for the same weight of materials, very materially better than others.

the

692. And you can only arrive at that result by actual experiment ?-In a way, yes; but, for

• See p. 236.

[Continued.

Mr. Murray-continued. instance, you can see that this cable which we have proposed for that long section is exactly the same as the Commercial cable. I said yesterday 40 words, but I see in this table it is 42 words, 80 one-third of that is 14 words, but I think we can calculate safely on 15, but we offer it to you as 12 words only.

693. The exact number of words can only be arrived at after the cable is made and tested and experimented on? No, we can determine it beforehand; I mean for our cables, and I dare- say naturally also the other contractors for their cables; but I mean to say you would be doing us an injustice if you simply said: here is a con- tractor who offers 522 lbs. of copper and 368 lbs. of gutta-percha, therefore his cable must be the quicker cable, if I may say so, than the one which you offer, which is only 500 lbs. of copper and 320 lbs. of gutta-percha.

694. What exactly do you mean by a word ? -A word is five letters. It is a conventional thing which is agreed upon in the cable world;

I mean it is always the same.

695. It always means five letters ?—Five letters, yes.

696. Does it mean five signals ?-Oh, no; I would advise you to ask Mr. Preece. I think it is only three signals that is the average of a letter, so that it is 15 signals to the word, is it not, Mr. Preece? (Mr Preece.) Something like that.

697. You have to make three signals to pro- duce one letter on the average ?--On the aver- age. Because the letters which are mostly used have got the shortest signal; they get only one signal or two signals; but the least frequently- occurring letters of course have more; 80 on the average I think it is three signals that are neces- sary. (Mr. Preece.) 37. The Witness.) 3·7, yes; so that it would be 18.5 signals to the word.

Chairman.

698. As regards this superiority you claim for your cables, do you attribute that superiority to the superior quality of the gutta-percha you use, or to the superior process of manufacture that it goes through in your works ?—No, it is the process of manufacture which does it.

Sir Donald Smith.

699. One word, to go back, about the tenders. We understood from you that when you tendered for the cable you did not look upon it as a direct tender, but that the Government desired to have some idea of what the cable would cost?—Yes; I gave practically, if you are comparing it, the same figures as are given in this little pamphlet.

700. And with regard to the proportion of business you supposed the new cable would take from the present cables: it was based on information available to every one, and you believed that the different Governments would be able to influence messages by the

G 4