66

The problem which has to be solved is one of considerable complexity, as will readily be realised when its essential characteristics are considered. A message handed in at the transmitting station has to be translated into a series of signals which can be telegraphically transmitted over a single telegraph wire. These signals, on arriving at the receiving station, must actuate a receiving mechanism in such a manner that a particular set of signals produces a certain definite movement of the mechanism; thus the signals corresponding to the letter a must cause the striking (or equivalent) of the typewriter key "a,' the signals corresponding to a notification of the end of a line must cause the shifting of the type-writer carriage ready for a new line, and so on. Now it is obvious that the signals as they are transmitted over the telegraph wire can only differ from each other by virtue either of their time arrangement or their magnitude. Each set of signals (corresponding to a letter) must be made up of one or more pulses of current, and one letter can only be distinguished from another by virtue of the pulses for the one being different in magnitude from those for the other, by their following one another at different intervals

of time, or by their lasting for different periods of time; of course, also, a combination of any two or of all three of these may be used. It is not possible for the telegraphic signals to be differentiated in space unless more than one wire is used to connect the two stations. It is equally clear that the distinction between the signals in their final form is one of space, and this is 1 "Setting Type by Telegraph." By Dona'd Murray. (Journal of the Institution of Electrical Engineers, vol. xxxiv., pp. 555. 1905).

so whether we consider the ultimate result, that is to say, the printed letter, or merely the alterations produced in the space relationship of the various parts of the printing mechanism which causes that mechanism instantaneously to print a particular letter. Thus we may say that what a type-writing telegraph has to do is the following:-it has to receive a message and translate it into a series of time or magnitude signals, to transmit these signals electrically over a wire, and to re-translate them into a series of space signals.

We have had occasion during recent years to describe several systems of telegraphy which aim at doing much the same thing as the Murray telegraph attempts, and it is of interest to compare the transmission methods used in these. Thus in the telautograph (see NATURE, vol. Ixiv. p. 107) the actual handwriting of the original message is transmitted and reproduced, and this is done by a combination of space and magnitude signals. Two wires are used, and current pulses of varying magnitudes sent along them which reproduce at the receiving end the motion of a pen at the transmitting station. Here the time element of the signals has no effect, and a letter is reproduced equally if it be traced in one second or in one hour. In the Pollak-Virag system (see NATURE, vol. Ixiv. p. 7) the telegraphic signals produce the motion of a beam of light which records in Roman letters the message transmitted. In this system the telegraphic signals differ from one another in their space relation and their duration. In the Murray system the signals differ from one another in their time relation.

We have pointed out that the first process is the translation of the message into a series of time signals, and for this purpose a time signal alphabet has to be chosen. Though this may at first sight seem a matter of secondary importance, it is in reality hardly too much to say thai upon the suitability of the alphabet selected will depend, more than upon anything else, the chances of success

FIG. 2.-Single-line Transmitter.

of the system. This fact has been thoroughly realised by Mr. Murray and others who have worked upon this problem, with the result that an alphabet has been finally devised which seems to possess in the greatest degree possible all the more important advantages. In it every letter or other signal which has to be transmitted is represented by a series of five time signals; the alphabet is therefore an equal letter alphabet," that is to say, each letter is composed of the same number of signal units (five in this case). The average number of units per letter is, al

corrections; these are made by punching five holes, thus blotting out all the holes already punched, this signal (of five holes) leaving the receiving mechanism unaffected. It is thus possible to wipe out any part of the message incorrectly written on the tape, and so produce a tape which will give an absolutely correct message when transmitted; this is facilitated by the fact that the operator

TRANSLATOR

PRINTER->

COLLECTOR

Н

55

-SPACE-SIGNALS

MAIN-LINE (TIME-SIGNALS)

SELECTOR

53

DISTRIBUTOR

PUNCHING MAGNET

VIBRATOR

SPACE-SIGNALS

RECORDER

59

can see the tape as it is perforated, letter by letter. The speed at which this perforator can be worked is about 120 letters (twenty words) a minute. The transmission can be carried on five or six times as rapidly, so that five or six operators working at these perforators can produce enough tape to keep the transmission line full.

is

The automatic transmitter shown in Fig. 2, and diagrammatically in Fig. 3 (collector). The tape is fed forward in the usual way by the star-wheel 15, passing across the end of an upright rod 1. This rod is pivoted as shown to the system of levers which oscillate about the centre 4, being kept in oscillation by the eccentric wheel 5, and making one oscillation for every unit on the tape. If this unit is a hole, the rod i enters this hole, the end 2 of the lever 2-9 is raised and the end g lowered, whereby the oscillation of the lever 3 brings the end 9 against the bar 11, thus pushing the contact lever 13 against contact 18. Here it remains until the next oscillation, and if this is the same as before, due to a second hole in the tape, it is not disturbed. It will thus be seen that successive signals of the same kind (either successive holes or successive spaces) are transmitted, not as intermittent, but as continuous signals. But if there follows a space in the tape the rod I cannot rise to its full height, the lever 2-9 is kept down at the end 2 and raised at the end 9, which comes in consequence against the rod 10 and forces the contact lever 13 over against contact 19, thereby breaking the punching current and sending spacing current into the line. The whole apparatus is driven by a phonic wheel motor in the usual way, the vibrating reed 23 sending currents alternately to the magnets 24 and 25, which keep the armature 26 in rotation. This is geared directly to the star-wheel 15, which has ten teeth, and is itself geared in the ratio of 10: I to the eccentric wheel 5, so that the latter makes, as already stated, one revolution for every unit of the tape. Now let us follow the message to its arrival at the

less than in an equal letter alphabet having a smaller average number of units per letter. Thus experience has shown that the actual average number of units per letter with the Morse code is only eight instead of thirteen. It must be remembered, also, that the Morse code is intended primarily for hand signalling, and consequently when time intervals are used the difference between any two which have to be distinguished manually or by ear must be fairly great. Thus the Morse dot consists of one unit, the Morse dash of three; were two units used for the dash instead of three, the distinction between the dash and dot would not be sufficiently marked. With machine telegraphy, on the other hand, there is no need to make such a great differentiation between the signals, as time intervals of one, two, three, and more units can all be distinguished, and in consequence it is possible to devise a shorter alphabet than the Morse code. It is not to be denied, however, that the use of a new alphabet is undoubtedly a disadvantage from the practical point of view, as it has to be learnt by the operators. This drawback is minimised by the fact that the operator does not print each signal separately as in operating a transmitting key; but it is nevertheless desirable, if not essential, that he should be able to read the message when printed on the transmitting tape.

To turn now to the apparatus used in the Murray system; the first operation, as in all automatic telegraph systems, is to punch the message to be transmitted on a Daner strip or tape." This is done by means of a keyboard instrument of the ordinary type-writer form shown, with the cover removed, in Fig. 1. On the tape will be noticed a double row of holes, which can be seen more distinctly in Fig. 4: the row of small holes serves only to feed the tape forward, both in this machine and in the transmitter: the larger holes are the signals punched in the tape. The actual perforator can be seen in front; it is worked by an electromagnet which

punches the necessary holes on the forward stroke and moves the tape one letter space (five holes) forward on its back stroke. On the right can be seen a lever which enables the tape to be pulled back letter by letter to make

tape by the punch 30, which is operated by the punching magnet. If the circuits of these two magnets are followed out it will be seen that both are controlled by the vibrating reed 34 in such a way that they operate alternately according as the reed is against contact 32 or 33. It will further be seen that the punching magnet is also controlled by the punching relay 27, the circuit being open in the position shown, and closed when the reed 41 is against contact 44, i.e. when punching current is coming through the main line and punching relay. It will be noticed at once that the distributor cannot work properly unless the tongue 41 of the punching relay is in synchronism with the reed 34. To obtain this synchonism is the object of the governing relay 28, which is cperated by the line current.

The

FIG. 5.-Murray Printer with Typewriter removed.

tongue of this relay vibrates between the contacts 42 and 43; when it is in contact with either the circuit of the vibrator magnet is closed, but during its passage from one to the other this circuit is opened. If this occurs whilst the contact 40 is open it can obviously have no effect on the oscillations of the reed, but if it occurs whilst this contact is closed it has the effect of diminish

intermittent current impulses to the spacing magnet due to the closing of contact 32. Line 3 shows the main line signals which, as pointed out in explaining the method in which the transmitter acts, are continuous and not intermittent. Line 4 shows the interruptions in the circuit of the vibrator magnet caused by the vibration of the reed of the governing relay which occurs at the beginning and end of every signal in line 3. In line 5 are the actual current pulses in the vibrator magnet due to the closing of contact 40. These are shown in step at the beginning. but gradually falling out of step, whereby, as will be seen, they are diminished by the interruptions shown in line 4, and are thus automatically brought back into step. The only remaining operation is to use the tape 45 (Fig. 3) to work either a type-writer or a type-setting machine. The Murray printer with the type-writer removed is shown in Fig. 5, and diagrammatically in Fig. 3. It will not be necessary to describe it in detail; the principle is that of the ordinary lock and key. The tape is fed forward letter by letter by means of the star-wheel 46; the reciprocating shuttle 47 carries a die block, which allows the five rods 48 to pass through the perforations in the tape when these are present. According as one or more of these rods passes through the tape, a particular set of slots in the combs, 49, attached to the rods is brought into line, the corre sponding lever 50 is pulled into the channel thus formed, and the corresponding type-writer key is depressed.

The complete set of Murray apparatus is shown in Fig. 6. On the extreme right is the perforator, next to it on the left the automatic transmitter, then on the same table the distributor in front and the relays behind. The translator and type-writer are on the small table at the left. We have only been able to give a brief description of the most important features of this very ingenious system; there are numerous points of detail which space does not permit us to describe. The system has been on trial for some time both in this

ing the duration of the current in the vibrator magnet. The reed 34 vibrates against two springs 36 and 37, so that its time of vibration is capable of great control by the magnitude of the current in the vibrating magnet. By setting it so that its natural speed is a little too high, it is possible by means of the controlling action of the governing relay for perfect synchronism to be obtained. The action will perhaps be more readily understood by the diagram, Fig. 4. This shows a piece of the transmitting tape at the top punched with the signals for the word Paris." In line 1 are shown the current impulses to the punching magnet due to the simultaneous closing of contacts 33 and 44. In line 2 are the regular

country and abroad, and has met with considerable success, it is now in use on several English lines. There can be no question after the perusal of Mr. Murray's paper that it possesses many advantages over its forerunners which should enable it to survive. It is stated that the automatic part of the apparatus can be run perfectly up to 200 words (1200 letters) a minute, but that no typewriter will stand the strain of being run at this speed, a maximum of 120 words being all that is allowable. It is. however, obviously possible to run the automatic part at top speed if necessary, and use two type-writers at the receiving end in the same way as at the transmitting end. MAURICE SOLOMON.

THE PERCY SLADEN EXPEDITION IN H.M.S. SEALARK. THE CHAGOS ARCHIPELAGO. OUR arrival in Mauritius on August 5 completed the first half of our cruise in H.M.S. Sealark, together with all our work directly connected with the Chagos Archipelago. This work may be divided under two heads, oceanography and biology. The former has been carried out mainly by Commander Boyle Somerville and his officers in view of the scientific objects of the expedition, but at the same time it is all of practical value for navigation in these waters. In many respects it has been of a singularly arduous nature; surveys by camping parties and deep soundings from the ship have been carried on simultaneously, together with numerous observations on the tides, currents, sea temperatures, &c. To a considerable degree it and all the work has been hampered by the heavy weather, which, contrary to all expectation, we have experienced, winds from south to east with heavy, confused seas, partially induced by the

com

paratively shallow waters of the Chagos Archipelago, and partially due to the current, which set in an easterly direction (against the wind) during the whole time we were in the group.

It is almost too soon to attempt to summarise any of the results of the cruise, but the soundings taken on our course from Ceylon to the Chagos and from the latter to Mauritius show that the archipelago is closely surrounded, both to the north and west, by the 2000-fathom line, and that there is at the present day no trace in the topography of the Indian Ocean of any former connection of the group with either the Maldives or the banks on the Seychelles-Mauritius line. The Chagos Archipelago appears, indeed, to stand by itself, being built up on a plateau rising to a depth of 800 fathoms in an ocean of an average depth of 2300 fathoms. Previously there were no bottom soundings between the banks and shoals of the group, but now a large series (more than 100) have been run, showing depths of 400 fathoms to 800 fathoms between the individual banks; from most of these a sample of the bottom has been obtained.

sand here and there. Indeed, our evidence points to the impossibility of any upward growth being in progress between the different Chagos banks, and to the probability of considerable current being felt even at 500 fathoms.

The reefs of the Chagos are in no way peculiar save in their extraordinary paucity of animal life, to which I referred in my last letter. Green weed, too, of every sort is practically absent. However, this barrenness is amply compensated for by the enormous quantity of nullipores (Lithothamnia, &c.), incrusting, massive, mammillated, columnar, and branching. The outgrowing seaward edges of the reefs are practically formed by their growths, and it is not too much to say that were it not for the abundance and large masses of these organisms there would be no atolls with surface reefs, &c., in the Chagos. The lagoon shoals of Egmont are covered by them, and alone reach the surface; having once done so they die and become hollowed out in the centre, finally resembling miniature atolls.

Main Pass

Pitt Bank

148

140

283 225

200

Centurion's r

150

Bank D 160

200

Broadly speaking, the Chagos group may be said to consist of three atolls to the north (Salomon, Peros Banhos, and Blenheim), the Great Chagos Bank in the centre (60 miles by 90 miles), and to the south two atolls, Diego Garcia and Egmont, besides certain submerged banks both to the north and south. Of these, H.M.S. Sealark has re-charted Salomon and parts of Peros Banhos, while Cooper and I have in addition examined the southern atolls. Salomon was very carefully surveyed, our intention being to make a comparison between its condition at the present time and when Powell's chart was made in 1837. The latter chart, however, proved to have been so carelessly | drawn that any close comparison is, I fear, useless, but the new chart should be of great value when it is possible to re-examine the atoll at some future date. Its section lines show that it arises in the last 400 fathoms by similar slopes to those of Funafuti, but it is a much simpler atoll, having only one passage, and more than half its reef crowned by land. Our numerous soundings and dredgings on its slopes leave no room for doubt but that its present reef is extending outwards on every side on its own talus, in fact, that the steep found round it (and, indeed, most atolls) is, in this instance, simply the slope at which coral and other remains from the reef above come to rest in the water. Its face was everywhere singularly barren; Lithothamnion, Polytrema, and, of course, reefcorals were not obtained below 50 fathoms. Further out, at 250 fathoms and over, the bottom was smooth and barren; the lead constantly failed to bring up any samples, while the somewhat broken and dented, but almost empty, dredges gave the idea of bare rock with a little muddy

130 100

Diego Garcia

مقد

FIG. 1.-Chart of the Chagos Archipelago.

In such a large group the conditions of the encircling reefs against the lagoons naturally vary very considerably. In general their inner edges reach the surface, and in the more open atolls the lagoon slope to 10 fathoms closely resembles the seaward slope. The bottoms of the lagoons are bare, rock, hard sand or mud, with shoals arising precipitously here and there, built up by a few species of coral, but largely covered by Xenia and Sarcophytum (as also are the only two submerged banks, Wight and Centurion, which we examined). Diego Garcia lagoon differs somewhat owing to its being almost completely surrounded by land. It has perhaps the most varied fauna in the group, and alone gives definite evidence of enlarging in every direction. Everywhere the land is entirely of coral origin. Diego Garcia shows signs of a recent elevation of a few feet, the present single island having been formed by the joining up of a series of separate islets on an elongated reef. The kuli or barachois (large shallow lakes) of the same island owe their origin to the same elevation, though elsewhere in the group they are generally due to

the successive washing up of beaches from the sea, enclosing areas of the reef. On the whole, there is singularly little change since the survey in 1837, and my impression is that Chagos has been for a long time an area of rest, and that the present condition of its reefs is mainly due to agencies still in action.

We have now examined the marine fauna in Salomon, Peros Banhos, Diego Garcia, and Egmont, and I would again lay stress on its comparative paucity and lack of variety as compared with the Maldives, Fiji, or Funafuti, though many of the forms are very common. In short, its general character is rather that of the temperate than of the tropical zone.

even

The land fauna is largely dependent on the flora, and the latter, except on small isolated islets and selected positions, has been destroyed to allow of cocoanuts being planted. The shores are everywhere fringed with Scaevola koenigii and Tournefortia argentea, both covered with a climbing bean. Behind these there was originally a forest formed of immense mapon (Pisonia capidia) and takamaka (Calophyllum inophyllum), with a few cocoanuts, Barringtonia, banyans, and other smaller trees, and an undergrowth largely consisting of immense Asplenium and other

brigge Fletcher has sorted the insects and finds about 110 species, most of which are probably indigenous; but the best season for the group would be in the rather hotter and damper north-west monsoon. On the whole, the land fauna and flora is much what one would expect to get, regarding the Chagos as a group of purely oceanic islands We expect to leave Mauritius toward the end of August for Cargados, Agalegas, and the submerged banks towards the Seychelles. Our cruise will be largely a dredging one, but the examination of Agalegas should be interesting. Meanwhile, Cooper and I hope to see some of the reels round Mauritius. J. STANLEY GARDINER.

IRON AND STEEL INSTITUTE.

FOR the first time the autumn meeting of the Iron and Steel Institute was this year held in Sheffield. An elaborate programme of visits to works and social functions was arranged, and no less than 1500 members and ladies were present, including members from all parts of the world. The opening meeting was held at the new university on September 26 under the presidency of Mr.

ferns and Psilotum, herbaceous dicotyledons being confined to the more open, dry, sandy, and stony parts; mangroves and Pandani are, curiously enough, not found. With the assistance of Dr. Simpson, we have collected the flora of each of the atolls, obtaining more than 600 specimens, about 140 species, of which probably only half are indigenous.

Of mammals there are only rats and mice, but there are traditions of dugong as well. Of birds the cardinal, sparrow, and mina have doubtless been introduced; noddies, frigates, and terns were breeding in enormous numbers on certain islands, though it was mid-winter; crab-plover, curlew, whimbal, and a sandpiper were common, and in the north-west monsoon buzzards, kites, and crows are said to be regular visitants. The green and shell turtles (Chelone mydas and C. imbricata) abound, the former coming on shore to deposit its eggs at night and the latter in the daytime. The only other reptiles are a marsh tortoise, perhaps introduced from Madagascar, and geckoes; there are no Amphibia. There is only one land shell, and arachnids and myriapods are scanty; the land crustacea are similar to those of the Maldives, but the coco crab (Birgus latio) is also abundant. Mr. Bain

R. A. Hadfield. Addresses of welcome were delivered by the Lord Mayor, the Master Cutler, the Vice-Chancellor of the University, Colonel Hughe (chairman of the reception con mittee), and by the president of the Sheffield Trades and Labour Council on behalf of the working men. Mr Hadfield, in reply, thanked the reception committee for the admirable work it had done, and gave interesting historical review of the Sheffield steel trade. Incidentally, he mentioned that the membership of the Iron and Steel Institute had now risen to 2200. After the reading of the minutes of the last meeting by the secretary, Mr. Bennett H. Brough. and the transaction of other routine business, the papers submitted wer read and discussed. In the first paper taken Prof. J. O. Arnold described the department of iron and steel metallurgy at the University of Sheffield. The main object borne in mind in designing the laboratory w the erection on a manufacturing scale of plant producing steel by the crucible, Bessemer, and Siemens processes.

Prof. J. O. Arnold and Mr. A McWilliam next contributed an important paper on the thermal transformations of carbon steels. For the research three steels were selected, saturated with 0.89 per cent. of carbon, unsaturated with 0.21 per cent. of carbon, and supersaturated with 1.78 per cent. of carbon. In the case of the unsaturated steel, the authors find that above Ar 3 (8102 C. the ferrite and hardenite are in mutual solution as a homogeneous mass. The Ar 3 change is accompanied by a segregation of the two constituents, which, if the oco!ing be slow, is probably completed in the Beta range temperature. After a fairly rapid cooling from 950° C the 0-21 per cent. carbon steel when quenched at 730° ( micrographically registered a segregation of ferrite so far advanced as strongly to suggest that such segregation mus! have begun at Ar3 and not at Ar 2. In other words. hardenite is insoluble in ferrite in both the Beta and Alpha ranges of temperature. It however still retains its identity as hardenite whilst falling through, say, 30° C. or 40 ( of temperature in the Alpha range, namely, from the end of Ar 2 at about 720° C. to the beginning of Ar 1 at about 680° C., at which latter temperature it begins to de compose into pearlite. The heating transformationsthis steel are substantially as follows:-At Ac 1 (abr 710° C.) in the Alpha range the pearlite begins to chang into hardenite, hence the carbide is soluble in the