ON AN IMPROVED CONSTRUCTION OF LOCK AND KEY.[12]
[12] By J. Beverley Fenby, Mechanical Engineer, of Birmingham.
Extracted chiefly from the Proceedings of the Institution of
Mechanical Engineers, 1866.
The simple _fixed-guard_ or _warded_ lock is so utterly worthless for
security, no matter what amount of good workmanship be bestowed upon it,
that it demands but short notice. It was contrived with the intention of
making the passage to the bolt intricate; but it will be seen at once
that this intricacy does not really offer any security. The wards of a
lock are circular arcs of thin metal, so arranged as to require a key of
peculiar pattern to pass amongst them, the shape of the cuts in the key
being a section of the wards. To make a really complicated box of wards,
and to cut keys which shall accurately fit their sweep, is a matter
requiring considerable manual dexterity; and some warded locks are
therefore expensive. But even with the best of them, all that it is
necessary to do for opening the lock is to take a blank key which will
properly fit the keyhole, coat it with wax, and then inserting it in the
lock, press it round against the wards, which will cause them to leave
an accurate impression of their section on the key. The parts impressed
are then cut out with small files, drills, and saws, and the occasional
use of fine cross-cut chisels. The key will then pass those wards which
impressed themselves upon it; and if these are the only wards, it will
go completely round and open the lock. If there are also other wards in
addition, not brought up flush with the first wards, the key is waxed
again and pressed against them, and then further cut out, as before.
This process is evidently one of absolute certainty, and the key so made
is in all respects as capable of mastering the lock as the original key.
These warded locks are however easily opened with merely a piece of bent
steel wire,--bent into such a sweep as will reach right round the wards
instead of passing amongst them, thus escaping all chance of being
obstructed by them. Such an instrument is called by burglars a “twirl.”
The fixed-guard or warded lock was the one in general use in the middle
ages.
The next kind of lock is the _tumbler_ lock, in which the bolt is moved
backwards and forwards by the key as usual, but these movements cannot
take place till a small lever with a stump on one side be lifted. This
lever and stump form the tumbler, which is held down by a spring; and in
the tail of the bolt are two notches, into one of which the stump fits
when the bolt is shot, and into the other when it is withdrawn. All that
is necessary to effect the picking of this lock is to lift the tumbler
high enough for clearing the stump out of the notch, and then draw back
the bolt. The tumbler may be lifted with one pick, and the bolt drawn
back with another; but generally one pick will suffice for both
purposes.
In the Barron tumbler lock the principle of double-action was
introduced.
The next improvement was the _lever_ lock properly so called, under
which designation the majority of the modern locks may be classed.
The Bramah lock was an admirable contrivance with remarkably beautiful
mechanism contained in a small compass; and since its invention there
have been several ingenious modifications of the same principle in
different radial locks, such as the Yale lock, in which the slides move
radially instead of axially. One advantage in these radial locks is the
greater difficulty in copying the keys, in comparison with the flat keys
of ordinary lever locks: this difficulty however is not an
insurmountable one.
A very ingenious addition was made to the action of the lever lock in
Newell’s American lock, which was shown in the 1851 Exhibition, and
described at page 89 of the present volume.
Though locks such as those already referred to exhibit great
dissimilarity of construction, yet there is one point in which they all
agree, and that is in the possession of a direct passage from the
outside to the works. Although various locks have been devised with the
object of having no direct passage to the works from the outside, one
consideration shows the inevitable existence of such a passage; namely,
that without it the key could not possibly at one and the same time
touch the hand of the operator and the works of the lock. It therefore
follows that any instrument which can pass in the same space as the key
may be brought to bear on the works, whatever may be their construction.
It can now be shown that, if picking instruments are thus brought to
bear on the works through the keyhole, there is a regular tentative
system whereby the picking of any lock with an open keyhole can sooner
or later be effected.
From the foregoing observations it is evident that there are two
important defects in the principle of the previous lever locks, which
being defects in principle are fatal to their security; namely, the
means of access to the works of the lock through the keyhole, allowing
of a series of attempts being made to open the lock by picking
instruments; and also the facility afforded for repeating the trial of a
false key made from a wax impression of the true key, and thus
perfecting it by successive alterations after trial. In consequence of
the possibility thus allowed of making these successive attempts either
by picking instruments or by a false key, it has been shown by the
cases that have occurred of locks of the best makes which have been
falsely opened, that, however numerous and complicated may be the
secondary impediments introduced into these locks, there can be no real
security against the ultimate success of sufficiently numerous and
persevering attempts, except by the adoption of some new principle of
construction specially meeting the above two defects.
[Illustration: fig. 53.]
In the invention of the Improved Lock and Key now to be described, and
which has been termed the _adytic_ lock,[13] the writer’s object has
been to meet this requirement. In fig. 53 is shown an elevation of this
lock, such as is made for an iron safe; two of the front cover plates
being removed to show the construction.
[13] From the Greek αδυτος, _inaccessible_.
The head B of the main bolt is of such a thickness as to be flush with
the face of the levers L and guard A; and the strap or tail D of the
bolt is thin, and passes behind the levers and guard, and also behind
the plate H H. The part of the tail D which would lie under the levers L
and cylinder C is removed, as seen in fig. 54, and replaced by a
separate flat plate or stump-bolt, carrying the stump S. This stump-bolt
has a projection K upon it, let into a recess in the tail D of the main
bolt, but with ¹⁄₁₀th inch vertical play in the recess. A spring in the
tail of the main bolt presses the stump-bolt downwards, keeping the
stump S in the notches of the levers L, as shown in fig. 53. The
stump-bolt can thus descend ¹⁄₁₀th inch at first without moving the main
bolt, and this amount of vertical movement is sufficient to carry the
stump in and out of the notches in the levers; but the stump-bolt cannot
descend further without taking the main bolt with it.
[Illustration: fig. 54.]
[Illustration: fig. 55.]
Immediately in front of the bolts comes the fixed plate H H, in which is
cut the cam groove shown by the dotted line J J; and also the vertical
slot E for the pin P to work through, together with another vertical
slot in which the stump S fits and works. This plate carries the centre
pin U on which the levers L turn. The levers are six in number, though
any other number may be used; and they occupy collectively ³⁄₈ths inch
thickness. In front of the plate H is fixed the guard A, which is made
of iron or steel, and has the brass cylinder C ground into it. The guard
is made a shade thicker than the levers L, in order to prevent the back
plate H and the corresponding front plate from being so tightened on the
levers as to impede their freedom of movement. The cylinder C is the
same thickness as the levers, excepting the centre boss F, which
projects from the back of the cylinder and works in a bearing in the
back plate H, and also projects in front through the thickness of the
two front cover plates. The small keyhole in the centre of the boss goes
only a short distance into the cylinder C, being merely for the purpose
of enabling the stem of the key M, fig. 55, to turn the cylinder; the
bit of the key is a separate piece, N, fig. 57, which is inserted
through a separate keyhole into the radial slot of the revolving
cylinder C, as shown at N in fig. 53.
This radial slot is cut in the side of the cylinder C that is furthest
from the levers when the cylinder is in the position shown in fig. 53;
and in the slot fits the slide block R, which is a steel block having a
pin projecting on each side. The back pin enters the guide groove J J in
the back plate H, as shown by the dotted line, and the front pin enters
the corresponding guide groove in the front cover plate, which is shown
removed. The back pin of the slide block projects through the back plate
H, as shown in fig. 56, and works in the cam groove O in the tail of the
stump-bolt S, fig. 54, which is so shaped that as the slide block
travels round the guide groove J J, shown by the dotted lines, it moves
the stump-bolt vertically as may be required according to the position
of the bolts and levers.
[Illustration: fig. 56.]
In the position of the lock shown in fig. 53, the bit N has been
inserted into the vacant space of the radial slot in the cylinder C, in
front of the slide block R. The size of this vacant space is ³⁄₈ths inch
long by ¹⁄₈th inch wide and ³⁄₈ths inch deep; and in the two front cover
plates of the lock, and also in the door to which the lock is attached,
a hole is made of the same shape. In the door there is no bearing for
the centre boss F, but only a small keyhole corresponding in size with
that in the boss F for inserting the stem of the key.
* * * * *
In the position of the parts shown in fig. 53, it will be seen that the
levers L are held pressing down against the circumference of the
cylinder C by their springs I bearing against the pin P. In this
position also the bolt spring between the main bolt and the stump-bolt,
fig. 54, presses the stump S down into the notches of the levers, so
that the levers are completely locked by the stump, as seen in fig. 53.
In order to unlock the lock, which in fig. 53 is shown with the bolt
shot, it is necessary that all the gatings G in the levers should be
brought precisely under the stump S. Through the centre keyhole F there
is no communication possible at any time with the levers L; nor will any
instrument, however slender, if passed into the radial slot through the
aperture at N be able to reach them, whether the cylinder C be in the
position shown in fig. 53 or turned round into any other position. For
the only difference made by turning the solid cylinder C is that the
radial slot in it is carried away from the aperture in the external
plates, and the solid part of the cylinder is brought opposite to the
aperture, which is thereby completely closed against the insertion of a
picking instrument. This construction accordingly not only precludes the
possibility of opening this lock with an ordinary key, in which the part
that acts on the levers is attached to the stem of the key, but it also
renders it an absolute impossibility to introduce a pick of any form, as
nothing can reach the levers L except a detached piece of such a size
and shape as to be capable of travelling round in the vacant space left
in front of the slide block R in the radial slot of the cylinder C.
[Illustration: fig. 57.]
For the purpose of unlocking the lock the bit N, fig. 57, is used. This
bit is of such a size as to fit into the vacant space of ³⁄₈ × ³⁄₈ × ¹⁄₈
inch in the radial slot of the cylinder C; and the indent at V is merely
for the purpose of ensuring the insertion of the bit in the right
direction, the external aperture for the bit being made with a
corresponding projection to fit the indent in the bit. This bit being
inserted through the aperture in the door, is pushed in by means of the
key stem M, which is flattened on two sides for that purpose, as shown
in fig. 55; and the bit is thus pushed home into its place in the radial
slot of the cylinder, as shown at N, fig. 53.
The key stem M is now inserted into the centre keyhole F, and the
cylinder is turned round by it in the direction shown by the arrow,
carrying round the slide block R and the bit N. The slide block R, while
moving through the concentric portion at the commencement of the guide
grooves J J, does not affect the bit; but by means of the cam groove O
in the tail of the stump-bolt, fig. 54, it moves that bolt so far as to
lift the stump S completely out of the notches in the levers L, which
are thereby left free to be raised. On continuing to turn the cylinder
C, the eccentric part of the guide grooves J J causes the slide block R
to move outwards along the radial slot, pushing the bit N before it; and
the bit is thus made to project beyond the circumference of the
cylinder, which it can then do, being no longer confined by the guard A.
The further projection of the bit as the cylinder revolves causes the
steps in the bit to lift their respective levers; and the steps in the
bit are so arranged that, when the cylinder arrives at the position
shown in fig. 58, all the gatings G are brought simultaneously opposite
the stump S, which is instantly shot down through the distance of the
¹⁄₁₀th inch play by the bolt spring. The bit N remains in contact with
the extreme part T of the levers while the stump S is entering the
gatings, the action of the bolt spring being so rapid that the bit
cannot move through any appreciable distance during the time.
[Illustration: fig. 58.]
In other locks a spring action of this kind would greatly facilitate the
picking, inasmuch as it would afford the gentle uniform pressure desired
upon the levers. In other locks, therefore, the bolt is caused to move,
and the stump to enter the gatings, by the direct contact of the key
with the bolt, instead of by a spring; but as the key, while moving the
stump into the gatings, is also altering its position under the levers,
a slight tremulous motion of the levers is thereby occasioned, which no
care in manufacture can obviate. This tremulous motion is aggravated by
the circumstance that, as the keyhole is open to inspection, it is
necessary to make all the levers fit flush with one another when down,
in order to avoid affording any clue to the shape of the key from the
positions of the levers; but as the various steps of the key, being of
different lengths, describe different arcs, the curves of the levers
when raised are of necessity in error to them all. The result of these
combined faults is that the gatings have to be made wider than the
stump, to allow a sufficient amount of play, thus introducing a fatal
element of insecurity in the construction of the lock, since the
security is of course enhanced in proportion as the gatings fit the
stump accurately. In the new lock, on the contrary, the arc T, fig. 58,
in each lever, can be shaped truly to its own proper radius, independent
of all the rest of the levers; and as the action of the stump is
instantaneous in catching the gatings as soon as they are all brought
simultaneously under it, the stump and gatings can be made to fit one
another with the most perfect accuracy, and without the slightest play.
On turning the cylinder C further round, the bit N passes from under the
levers, which remain held back by the insertion of the stump in the
gatings; and just before reaching the position shown in fig. 59, the
slide block R has pushed the bit completely out of the radial slot, and
the bit falls down as shown in fig. 59, and drops through a hole into
the inside of the safe that is locked. At this point the back pin of the
slide block comes in contact with the lower side of the cam groove O in
the stump-bolt, fig. 54; and by turning the cylinder C onwards to the
position shown in fig. 60, the withdrawal of the bolt B is completed,
bringing the parts into the position shown in fig. 60. In these drawings
only one lever L is shown; but there are altogether six levers, as
shown in the sectional plan, fig. 56. The pin P is fixed in the tail D
of the main bolt, so as to travel with the bolt; and by this means the
springs I are released from strain, as shown in fig. 60, as soon as the
bolt is withdrawn.
[Illustration: fig. 59.]
* * * * *
From the nicety with which the various parts of this lock are
constructed, it is evident that the levers must be very accurately
lifted by the bit of the key in order to withdraw the bolt; and
therefore any error in the bit, such as would occur with a false bit,
will effectually prevent the lock from being opened. This may be
illustrated by supposing the false bit to be so close an imitation as to
have five of its steps absolutely correct, and the sixth only slightly
wrong: though it is almost impossible that such a near approach to
correctness could be attained in practice. The counterfeit bit being
inserted in the lock, and the cylinder turned round, all will go on the
same as with the true bit, up to the time when the false bit reaches
the point T of the levers, as previously shown with the true bit in fig.
58. Here a change of action takes place; but what is the nature of the
change the operator has no means as yet of ascertaining. In the case
supposed, where five of the steps in the bit are right, but the sixth is
wrong, the gating of the sixth lever does not precisely coincide with
the others, nor with the stump S; and the consequence is that, at the
critical moment when the stump ought to spring into the gatings and hold
back the levers from falling forwards, it will be prevented from
entering the gatings, owing to the entrance being partly blocked up by
the one lever, which stands more or less across it.
[Illustration: fig. 60.]
The fact, however, that the stump cannot enter the gatings, does not
become known to the operator until the cylinder C has been turned
further round, so as to bring the slide-block pin in contact with the
lower side of the cam groove O in the stump-bolt; and before this point
has been reached the false bit has already passed clear of the levers,
which, not being retained by the stump, are instantly thrown forwards
again by their springs, and locked in their original position by the
stump entering the notches. At the same time the false bit has dropped
into the inside of the safe in the same manner as the true bit, as shown
in fig. 61.
[Illustration: fig. 61.]
Hence a person putting a false bit into one of these locks will not only
infallibly lose it at the very first trial, but will do so without
gaining any information as to the nature of its inaccuracy; for as the
gatings of the levers cannot be seen or felt, all that can be told about
the action of a false bit is, that it has failed to open the lock. In
fact, a counterfeit bit passes under the levers, and through the lock,
just like the true bit; and it is only the stoppage afterwards met with
of the bolt that indicates the failure of the false bit, which is by
that time gone beyond recovery. Whatever amount of labour, therefore,
may have been spent on the fabrication of a counterfeit bit, this bit
can only be tried once, so that no alteration can afterwards be made in
it.
Nothing that can be inserted into the radial slot of the cylinder C
through the aperture in the front plates can do any injury to the lock;
and a charge of gunpowder inserted in that way would only blow out
again at the orifice without damaging the lock, both the apertures for
the key being merely blind holes with parallel sides.
* * * * *
[Illustration: fig. 62.]
[Illustration: fig. 63.]
For the manufacture of the bits for the keys of this lock a self-acting
machine is employed, in which the height and width of the several steps
in the bit are regulated by adjustments of very great accuracy, and
admitting of an almost endless variety of figure for the bits. This
key-cutting machine is shown in figs. 62 to 64, and consists of a small
circular saw A running vertically, of the same thickness as each step
in the bit I, which is brought up to the saw by the slide-rest B. The
bit I is fixed in the holder C, which rocks upon a centre, so as to give
the required curvature to the edge of each step in the bit when cut by
the saw, as shown in the full-size section of the bit-holder, fig. 63.
The adjustment of the depth of cut is effected by the set screw D upon
the slide-rest coming up against the eccentric ring E upon the bed of
the slide-rest; this ring is turned round by hand, and set to sixteen
different positions by means of the catch-pin F and the sixteen holes on
the circumference of the ring, allowing of sixteen different depths of
cut. The lateral adjustment for the pitch between the successive steps
of the bit is effected by the two bed-screws G G acting on the
slide-rest B, having a dividing plate on the head, and such a pitch of
thread that one turn of the screws traverses the slide-rest through the
exact distance of one step in the bit. The occurrence of any play or
backlash is entirely prevented by having the screws placed one at each
end of the slide-rest; so that by slacking back one screw through one or
more turns, and then advancing the other through the same number of
turns, the slide-rest is always held with perfect steadiness between
them, filling exactly the space between the ends of the two screws.
The number of changes admissible in this key-cutting machine, if used
for making keys for locks having six levers, is the number of
permutations that sixteen terms are capable of when taken six together,
which is upwards of sixteen millions. Some of these changes are so
slight that too great accuracy of workmanship would be required to make
the locks accordingly; but of those changes that differ from one another
so far that no lock could be opened by any other than its own key, more
remain than could be used up by all the locks in the world.
[Illustration: fig. 64.]
The writer may observe that it was the study of the circumstances of the
great gold robbery on the 15th of May, 1855, by Agar and his
confederates (when two of the best lever locks were picked, and gold
stolen weighing upwards of 200 lbs., while in transit on the
South-Eastern Railway from London to Paris, packed in three sealed
iron-bound boxes, inclosed in a bullion safe, secured by those locks),
and of the various modes of picking locks, which led him to turn his
attention to the achievement of what had been so long and perseveringly
sought after, namely, an unpickable lock. The principle of a detached
bit has been previously tried, in so far as that locks have been made in
which the bit of the key was deposited in the lock by unscrewing the key
stem, and then withdrawn by screwing in the stem again. But inasmuch as
the detached bit, even though it failed to open the lock in the case of
a counterfeit key, could always be brought back again to the keyhole and
removed, this admitted of a repetition of attempts with successive
alterations of the one counterfeit key, without the certainty that any
warning would be given by the lock of such attempts having been made.
In another still more complicated lock with a detached bit there were
two keyholes, into one of which the bit of the key was put, and the stem
being then unscrewed from the bit, was put into the second keyhole and
turned round so as to close the first keyhole over the bit; a separate
handle was then turned to work the lock, six separate operations being
required for either opening or closing the lock. Further, a kind of
retainer has been attempted by so arranging the lock that, if any key
was put in but the right one, it was held in the keyhole in such a
manner that it could never be got out. In this case, however, if the
false key would not open the lock, neither would it let even the right
one do so, and it would be necessary to break open the door secured by
the lock.
* * * * *
In the new lock here described, the special points that have been aimed
at are the following:--
Firstly, in no position of the lock is there any access to the works
from the outside through the keyhole. This access through the keyhole is
more or less a defect in all other principles of lock, as it admits of
feeling and manipulating the works for the purpose of getting
information for picking the lock in the absence of the right key;
whereas in the new lock there is no opening whatever at any time, except
the two plain parallel recesses into which the key and the bit are
fitted. From the moment the turning of the lock commences both these
recesses are effectually blocked up: the one for the bit being conveyed
bodily away from the keyhole, and its place taken by the solid metal of
the cylinder; while the other is completely filled by the key, which
cannot be withdrawn except by turning it back to the original position.
In consequence of this construction no injury can be done to the lock by
explosion of gunpowder in the keyhole, the only openings from the
outside being parallel at their sides, and not communicating with any
portion of the interior of the lock; and the simplicity and solidity of
construction are such that the revolving cylinder is made practically
air-tight within its bearing. This effectually prevents all attempts to
open the lock by picklocks, and leaves no alternative but the attempt to
make a sufficiently accurate copy of the true key.
Secondly, as no clue whatever can be obtained from the outside of the
lock respecting the key required, the attempts upon the lock are thus
limited to the chance of obtaining a wax impression of the true key. The
difficulty of making a counterfeit key sufficiently correct by this
means for opening one of the best of the previous constructions of lock
is very great; but in the new lock this difficulty is greatly increased
by the fact of the levers remaining absolutely stationary while the
stump enters the gatings, in consequence of which the gatings are made
so close a fit to the stump that an exceedingly minute error in the
lifting of any of the levers is sufficient to prevent the lock being
opened. This extreme delicacy of construction can be carried out
practically without objection in the new lock, because there is no
possibility of putting a strain from the key upon the stump, so as to
cause injury by forcing it at the moment of entering the gatings; for
the only force acting upon the stump at that time is the uniform
pressure of its own spring. In addition to this source of increased
safety, there is the still more important circumstance that only a
single trial can be made of each counterfeit bit; because, if carried
forwards far enough to try its effect in opening the lock by passing the
levers, the bit is inevitably lost by falling through the lock and
inside the door. Thus not only is all chance prevented of a second trial
with the same key, but the bit retained inside the door gives warning of
the attempt having been made, and shows how near the counterfeit key has
approached to the original. The numerous cases that have occurred of
attempts to open locks by counterfeit keys, such as the remarkable
instance previously referred to, show that even with the most practised
hands it is next to impossible to make from a wax impression a key that
will serve for opening a good lock the very first time it is tried; and
the striking importance is therefore seen of this arrangement in the new
lock, which prevents more than a single attempt being made with a
counterfeit.
Thirdly, another advantage to be named in this lock is that the stem
alone of the key is required to lock it, but it can only be unlocked by
the complete key. The stem, therefore, can be left by the principal of
an establishment for locking up by a subordinate; but the bit, which is
the essential part of the key required for opening the lock, need never
be used or seen by any one but the principal himself. As the hole in the
external door-plate for the stem of the key has a notch on one side only
to admit the key stem, and the cylinder is prevented from making a
complete revolution, the stem of the key cannot be withdrawn from the
lock except when the bolt is shot; so that its absence from the keyhole
serves as a proof that the bolt is shot.
Fourthly, one other advantage in this lock is its simplicity and
solidity of construction. It contains no more parts than the simpler
forms of lever lock having the same number of levers, and the total
number of separate pieces in the complete lock is only sixteen. The
principle of security, therefore, upon which the new lock is
constructed, avoids entirely the complications and the delicate and
minute class of work rendered necessary in other locks by the use of
detectors and the other auxiliary contrivances employed for increasing
the difficulty of picking.
* * * * *
Mr. Fenby exhibited, at a _conversazione_ of the Institute of Civil
Engineers, and at the meeting of Mechanical Engineers, specimens of his
adytic lock, and showed its action both with the true key and with
counterfeit keys; and he showed by trial that the counterfeit failed to
open the lock, notwithstanding that by means of the permutating cutting
machine it had been made a much nearer approach to a perfect copy than
was practicable in the best handwork from a wax impression. He also
exhibited the key-cutting machine employed for cutting the bits; and
also a set of burglar’s tools employed for drilling into the door of an
iron safe sufficiently for breaking open or removing the lock, showing
that the hold required for giving the cutting pressure upon the powerful
drill employed for the purpose was obtained by a steel cross piece
inserted into the keyhole and turned at right angles, so as to hold
across inside the lock; but in the new lock, as the keyhole had no
opening into the lock, and only a slight shoulder on one side, no means
were afforded for obtaining the required hold for the drill.
The following are the salient points of the discussion that followed the
reading of his paper:--
The Chairman remarked that the paper just read gave a very excellent and
clear description of the detailed working of the new lock, and he
thought this construction of lock was a most valuable one, as affording
real security against all fraudulent attempts. He inquired whether there
would be any possibility of tampering with the lock by examining it upon
the inside of a safe door, whenever the door might happen to be left
unlocked.
Mr. Fenby replied that there was no means of tampering with the lock
from the inside of the door, as the two keyholes for working the lock
were only in the front face of the door, and the lock was all closed up
on the inside of the door, excepting the hole through which the bit was
allowed to drop out; but this would be useless for the purpose of
tampering with the lock, as the bit dropped down a tube leading to the
bottom of the door, through which no examination of the lock could be
successfully made.
The Chairman inquired whether there was any provision against the bit
being accidentally locked up inside the safe, in which case it appeared
the lock could not be opened again.
Mr. Fenby replied that the owner of the safe must of course be careful
after unlocking the safe to take the bit out before locking it again,
otherwise there would be no means of opening the lock afterwards with
that key. As a precaution, however, against any such accident, each lock
was provided with three bits, all duplicates, one of which would be kept
in the pocket for use, while the two others would be preserved in a
place of safety for the chance of any such contingency. Moreover, in
most of the safes fitted with these locks, the tube through which the
bit dropped had been made of such a length as to carry out the bit on
opening the door, dropping it into a small tin tray outside the safe;
and by this means the accidental locking in of the right bit was
rendered impossible. One of the advantages of the new lock was that the
stem of the key was not required to be kept constantly in the possession
of the owner, but it might be left in the lock, as the bit alone was the
valuable part of the key; and as the bits were of such small size and
convenient shape, a number of them might readily be kept in the pocket
by a person having charge of a number of safes, without the
inconvenience attending a large bunch of ordinary keys. In the case of
an attempt being made to open the lock with a counterfeit bit, the
advantages of retaining the counterfeit inside the safe were not merely
that the person attempting the lock was deprived of his instrument,
while the proprietor immediately discovered the attempt upon the next
occasion of opening the safe; but the retention of the counterfeit
itself afforded the means of judging, by a comparison with the true bit,
whether the attempt had been made altogether in the dark as to the
actual construction of the lock, or whether it was likely that some clue
regarding the true bit had been obtained by means of a wax impression or
otherwise. In the latter case the owner of the safe might think it
desirable to have the lock taken off, and the arrangement of the levers
altered, and a new bit made so as to baffle any further attempts.
Mr. W. S. Longridge observed that the inconvenience that had been
alluded to with the new lock, of accidentally locking up the bit inside
the safe, was no greater than occurred with an ordinary safe lock if
ever the key was accidentally lost; in either case, unless the
precaution was taken of keeping a duplicate in reserve, it would of
course be necessary to have the safe broken open.
The Chairman inquired how the ideas had been arrived at of separating
the bit from the key, and of preventing all access to the works through
the keyhole, and also of retaining the bit inside the door after any
attempt at unlocking.
Mr. Fenby replied that his attention had in the first instance been
attracted to the subject of the picking of locks as a mechanical
problem, and he had found that there had hitherto been no principle in
lockmaking which could effectually baffle persevering attempts at
picking. For although there were certain complicated constructions of
locks, having many points of excellence, they had all yielded in time to
the picking instrument in clever hands; and it must be remembered that
any individual lock when once constructed remained stationary as
regarded subsequent improvement, whereas the art of picking that lock
was continually progressing towards success, with all previous
constructions of locks, and it was clear therefore that the lock must
ultimately be defeated. He had been further stimulated in the
investigation of this subject by the occurrence of the great gold
robbery referred to at page 188; and the circumstance which had struck
him most forcibly in connection with that robbery had been that locks of
the best make hitherto known had admitted of seven successive trials
being made upon them without detection, each trial furnishing the
information for further perfecting the counterfeit key, until the locks
were at length opened.
These considerations had led him to the conclusion that two points were
established and were required to be kept in view for the construction of
any lock that should be really secure against fraudulent attempts. The
first point was that wherever a man could get instruments into the lock
he could ultimately solve any problem laid before him by the maker of
the lock, as the lock when once made could be tried any number of times
if an instrument could be got into it at all. Hence he had concluded
that it was requisite for all access to the interior to be cut off, so
as to preclude all possibility of getting a pick-lock in; and this was
accordingly accomplished by adopting the plan of separating the bit from
the stem of the key. The second point established was that it was
necessary to prevent the possibility of making a succession of trials
with the same counterfeit key; and it had then struck him that, if the
bit of the key were arranged to drop inside the safe in unlocking, there
would be no means of going on gradually improving and touching up the
counterfeit from the results of previous trials, as the false bit would
be irrecoverably lost in the very first attempt, without furnishing any
clue whatever as a guide for alteration in a subsequent trial. The first
lock that he had invented for meeting the requirements thus pointed out
had been made with a solid block having a tunnel through it, but
involving the same principle of retaining the bit of the key and keeping
the levers inaccessible from the outside. Subsequently, however, he had
abandoned that construction and produced the new lock shown in the
drawings, having the revolving barrel with radial slot.
The Chairman proposed a vote of thanks to Mr. Fenby for his paper, which
was passed.