October 12, 191 1] 



NATURE 



4»9 



button, which operation again earths the vertical line. 

 VR is operated, and spring r making contact with 5 

 puts earth on the left-hand side of GM by way of 

 switch-lever E (position 3). K also is operated by 

 its upper coil, and the closing of its springs e and d 

 brings the battery on the right-hand side of GM. 

 The movement of GM's armature puts the generator 

 (which is running continuously) on to the wipers, and 

 thence to the required subscriber's (Z) loop, across 

 which is placed his magneto bell, in series with a 

 condenser. When Z replies, the current circulates 

 through springs c. j, through the upper coil of J, 

 round the loop, through J's lower coil, switch-lever F 

 (position 3) to earth. J's two coils are now assist- 

 ing each other, and spring b moves over to c. As b 

 is joined to K's lower coil, this movement takes the 

 battery from K's lower coil and puts on earth in- 

 stead. The current now passes through K's upper 

 coil, through VR, round A's loop, through RR and 

 K's lower coil to earth via c. The currents in K, 

 however, are opposing each other, so that the arma- 

 ture is not affected. The two subscribers are now 

 "through," as in Fig. 6. The path traversed by the 

 current on the left is shown bv the dotted lines, and 

 on the right by the heavy lines. The arrows through 

 J and K show the direction through the separate coils, 

 and the shading indicates the energised relays. The 

 small skeleton diagram in Fig. 6 shows how the 

 speaking circuit is made up. On Z's side the two coils 



Fig. 7. — Clearing. 



ol J serve as impedance coils, whilst K's upper coil, 

 plus VR and K's lower coil, plus RR, serve the same 

 function on A's side. 



On the conclusion of the conversation both sub- 

 scribers hang up their receivers. In its passage down- 

 wards the switch-hook (momentarily) earths both the 

 vertical and rotary lines simultaneously. The result 

 of this action on A's part is that his rotary line is 

 earthed at both ends. RR and the lower coil of K are 

 thus short-circuited. K is then actuated by his upper 

 coil, and by the movement of e from f to d current 

 is cut off from J, which, ceasing to be energised, 

 allows its armature to fall back. This causes b to 

 leave c and return to a, substituting the battery for 

 earth on the lower coil of K. Current now passes 

 through both coils of K in the same direction, and 

 thus continues to energise K, whilst both VR and RR 

 being actuated, their outer springs make contact with 

 each other and put earth on the right of the coils 

 of CM, the clearing magnet. The other side of 

 CM being connected with the battery through d e, the 

 clearing magnet is energised. This is shown in 

 Fig. 7, where the dotted line indicates the circuit of 

 CM. The shaft is thus restored to its normal position 

 and the circuits are cleared. 



When a second caller attempts to get through to 

 another subscriber who is alreadv engaged, the follow- 

 ing action takes place. As shown in Fig. 6 the earth 

 on the left hand of J's lower coil is made bv the lever 

 F of side-switch in its third position. This puts earth 



NO. 2189, VOL. 87] 



on the "engaged " wiper, and through it on to all Z's 

 contacts in the busy multiple. The new caller turns 

 his disc twice in the usual way, thus getting on to 

 Z's contacts in the line multiple. His (the new- 

 comer's) side-switch, however, is still in position 2, 

 so that, although his line wiper is on the required 

 contacts, the wiper itself is still isolated. The final 

 current over his rotary line causes RR and SM to 

 be actuated. A contact on the latter (shown only in 

 Fig. 3) is closed. The battery is already connected to 

 the left hand of the newcomer's clearing magnet by 

 means of his springs d and e. The closing of the 

 contact on SM put the right-hand side of CM through 

 to the earthed busy bank by way of lever F (position 

 2) of his own side-switch. His clearing magnet thus 

 acts and his lever L returns to zero. By another 

 contact (not shown) the busy signal from the generator 

 is given over the newcomer's line, advising him that 

 the required subscriber is engaged. 



From the foregoing brief sketch it is hoped that the 

 principle of working may be seen : the actual arrange- 

 ments in practice involve large modification and ex- 

 tension. As only a certain percentage of the switches 

 are in use at the same time, it is easily seen that it is 

 unnecessary to provide one for every subscriber. A 

 much simpler piece of apparatus, the "line switch," 

 is therefore substituted, and only a comparatively 

 small number of switches proper provided. The func- 

 tion of the first is simply to put a calling subscriber 

 through to a dis- 

 engaged switch — now 

 slightly modified and 

 called a " selector " — 

 by which a certain 

 group is selected. The 

 A subscriber having got 

 through to the required 

 group, now utilises a 

 second switch, arranged 

 practically as we have 

 described and termed a 

 "connector," through 

 which he obtains access 

 to the required corre- 

 spondent. A still larger 

 exchange will require a first selector, a second, &c, 

 selector, and finally a connector. Taking the common 

 case of an exchange with a maximum of 9999 sub- 

 scribers : the line switch puts the caller through to a 

 first selector, by which the thousands digit is selected. 

 The hundreds figure is then picked out by the second 

 selector, and the tens and units by the vertical and 

 rotary movements of the connector. 



Arthur Crotch. 



THE SUMMER OF 1911. 

 "THE summer of 191 1 has been remarkable in so 

 -*- many ways that without doubt it will receive 

 the special attention of meteorologists, and will in 

 course of time be dealt with very thoroughly, as it 

 well deserves to be. Having for many years past 

 kept touch with the published Greenwich weather 

 records, a comparison of the present summer with the 

 observations of the past seventy years, from 1841, may 

 be of interest from one not officially attached to the 

 Royal Observatory. 



The exceptional character of June, July, and August 

 lead naturally to the supposition that the summer 

 proper, as limited to the three months, would beat all 

 previous records in many ways, and this impression 

 is amply supported by weather statistics. 



The summer six months, April to September, can 

 also claim a record so far as temperature is con- 

 cerned. The mean temperature for the six months is 



