G-50 10/7/44 RS#42937)
Op-20-G-md
(SC)A6-2(8)
From: OP-20-G-1
To: OP-20
Via: OP-20-G
Subj: Review of Activities and Future Plans at the U.S. Naval Computing Machine Laboratory, Dayton, Ohio.
Ref: (a) Comdr. Wenger’s memo to DNC, dated 30 May 1944, under cover memo dated 31 May 1944.
(b) VCNO’s memo for Chief Naval Personnel dated 11 November 1942.
(c) Comdr. Wenger’s memo to OP-20, dated 10 May 1944.
1. The U.S. Naval Computing Machine Laboratory, Dayton, Ohio, was set up by a request of the Vice Chief of Naval Operations dated 11 November 1942, reference (b).
2. The primary function of this Laboratory was the design and construction of the Bombe equipment necessary for the solution of German cipher systems. The outstanding points in the history of the Bombe project are contained in reference (a). The purpose of the present memorandum is to formulate the present status of plans for the activity.
3. The major projects carried out by the Laboratory at Dayton were the original Bombe development which included the production of ninety-six (96) units and the contract for fifty (50) additional units which was initiated at the request of the British in March, 1944. The latter contract was cancelled, after twenty-five (25) units had been produced, at the direction of the Director of Naval Communications with the concurrence of the British.
4. Because of the introduction by the Germans of a pluggable reflector in the Enigma machines used on certain German Air Force circuits, a project for the construction of a special machine, called the DUENNA, for the solution of this system was placed at the Laboratory. The present status of the DUENNA project is as follows: one unit has been completely assembled and is undergoing operating tests. The second unit is in assembly. Materials for five (5) units are at hand and certain critical items for a total of ten (10) units are on order. The distribution of this pluggable reflector by the Germans has not been as wide as was expected. Unless the European situation changes rapidly, two DUENNA’s should be sufficient.
5. The primary function of the Bombe equipment is the determination of the plugging and wheel setting at which an assumed crib will produce a given cipher text. The DUENNA function is similar with the additional factor of determining the required reflector plugging. The fundamental cryptanalytical principles, of which the Bombe is an electro-mechanical representation, were used by the British for three-wheel Enigma machines, and were discovered independently by American cryptanalysts. The cryptanalytical principle behind DUENNA was the subject of research by American cryptanalysts beginning about June 1943, and had been successfully established before the Germans introduced the pluggable reflector. Although the fundamental principles involved in both Bombes and DUENNAS have been firmly established, varied usage by the Germans has necessitated intricate modifications on certain units in order to handle special problems quite different from straight cribbing of special messages. These modifications have been carried out partially at the Dayton Laboratory and partially by the maintenance force at the Naval Communications Annex. A fundamental distinction between American and British equipment is in the number of Enigmas contained in a Bombe unit. British machines contain thirty-six (36) Enigmas while the American machines contain sixteen (16). The decision to restrict American Bombes to sixteen (16) units was based on calculations which established that a sixteen (16) letter crib was sufficient to determine the unknown factors in the problem and hence the use of more than sixteen (16) Enigmas was wasteful of equipment under normal circumstances. The sixteen (16) unit Bombe proved extremely effective for a large percentage of German cipher problems.
6. Several German cipher systems which were intercepted by our own stations had never been solved by the British, notably the SEAHORSE traffic which is a Naval Attache circuit between Berlin and Tokyo and the SUNFISH traffic which is used by German units including blockade runners operating in transit or in the Far East. Cryptanalytical work leading to the initial break of both these systems was done by Op-20-G personnel. The solution of these systems is not one of cribbing but depends upon the knowledge that a message setting has been enciphered twice in the message. This problem of double encipherment can be handled by Bombe techniques but two Enigmas are necessary where one (1) suffices for a normal crib. Thus, these indicator or “throw-on” problems could not be handled on our sixteen (16) unit equipment. The initial runs on this traffic were made by the British at our request. However, British emphasis on European problems left them little time for this traffic, and hence it was decided to combine two (2) of our Bombes into a double unit. This was done in spite of British coolness to the proposal. The Laboratory was also instructed to combine two (2) of the additional fifty (50) Bombes into a second double unit. This course of action appears to have been justified because our one (1) machine, since going into operation seven (7) months ago, has broken out all the SEAHORSE traffic read seventy-two (72) days since then, and thirty-three (33) of the forty (40) days of SUNFISH read since then. At the present we are reading SUNFISH traffic solidly, but at the expense of SEAHORSE which was being read several weeks ago.
7. In addition to the double unit modification, many special attachments for making Bombes more generally applicable have been carried out. Most important of these are the following:
a. STANDARD GRENADE.
Grenades are attachments to Bombes which determine the setting when the plugging has been recovered. A four-letter crib is sufficient for this purpose. Four (4) units are in operation.
b. SLIDE GRENADE.
This device cuts the Grenade running time for Enigmas with rotatable reflectors by a factor of 26. They are of great interest at the present time because the recently captured Japanese Enigma has a rotatable reflector. Two (2) of the Slide Grenades have been completed.
c. DRAG GRENADE.
This Grenade applies a short crib to thirteen (13) positions simultaneously. It proves particularly useful when no crib is available, because such trigrams as “EIN” and “SCH” usually occur somewhere in the message. One (1) Drag Grenade is available.
d. PARALLEL GRENADE.
The Parallel Grenade provides four (4) standard Grenades which can be run simultaneously on a Bombe and thus uses the Bombe to full efficiency. This is of particular interest because the recently captured Japanese Enigma machine has five (5) notches on each wheel and therefore will be best handled by the Parallel Grenade. One (1) Parallel Grenade is available.
e. COAST GUARD GRENADE.
A special Grenade constructed for the solution of double encipherment problems occurring on certain clandestine circuits. This unit is near completion.
f. CILLI GRENADE.
This is a special Grenade which makes possible simultaneous running of several wheel orders.
g. SELF-DETECTION CIRCUIT.
This circuit was designed to eliminate certain extraneous stops called “boxing” stops occurring on indicator problems. Four Bombes and the double unit have been equipped with this modification.
h. SQUELCHER CIRCUIT.
This is a special circuit to eliminate stops which enables weaker cribs to be handled. This has proved particularly effective on certain paired day recoveries of submarine traffic.
i. INVERTED MACHINES.
It is found necessary to convert the wiring on certain Bombes in order to take advantage of turnover positions. These have been particularly useful for paired days on submarine traffic in which there is an unknown stecker but the wheel order and rings are known. Eight (8) machines are available.
j. PLUGGABLE REFLECTORS.
The reflector wiring on the Bombes is contained in a double-ended Jones plug. This arrangement permits change of reflector wiring but not rapid enough to cope with the probably German use of the pluggable reflector. To make rapid reflector changes possible, all Bombes are being provided pluggable reflectors.
k. DOUBLE UNIT
This modification was introduced on twenty-five (25) Bombes in order to be able to run certain weaker cribs.
8. In addition to the Bombe equipment, hand machines for secondary testing and breaking out Enigma traffic have been built. The following are some of the more important items:
a. M-9.
A hand machine for testing results obtained from the Bombe. Approximately sixty (60) are on hand.
b. M-8.
Modified EOM’s to carry out the Enigma decipherment from punched tape. Eight (8) machines are on hand.
c. BOA
A pluggable reflector for use in the M-9 machines. All M-9 machines have been equipped and ten (10) special units have been supplied for use with DUENNA.
9. Although the primary activities at the Laboratory at Dayton have been directed to the development of equipment for the solution of the Enigma machine, several projects have been placed at Dayton for machines aiding the solution of subtractor ciphers and other systems used by the Japanese. These projects are outlined as follows:
a. COPPERHEAD I.
COPPERHEAD I is an electronic tape passing device which lines up subtractor text in depth on the basis of double-repeat occurrences. This double-repeat search when carried out by IBM equipment is extremely laborious. The COPPERHEAD equipment will do a complete search on the two weeks’ volume of JN-25 traffic in about twenty-four (24) hours. These machines are now practically complete; delivery is expected within two weeks.
b. MAMBA.
MAMBA is a device for lining up cipher messages with respect to recovered portions of the additive book, a process which is now running by hand. MAMBA will do the same job with a great saving of time and personnel. Delivery of this equipment is promised within six (6) weeks.
c. RATTLER
The RATTLER is a device which determines the settings of Japanese 157 cipher machine on the basis of a four (4) Kana crib. Modifications of the RATTLER which are now in operation are being made so that it may be applicable to the Japanese Naval Attache machine. The extended use of the 157 cipher machine appeared imminent in July when three (3) additional RATTLER units were ordered. There has been no evidence to date of the expected wider usage so that the RATTLER program is currently being reviewed.
d. VIPER
VIPER is a representation of the Japanese 157 machine and is used in the decipherment of traffic after the keys have been recovered. Seven (7) of these units have been completed; two (2) have been shipped to Honolulu and two (2) to Melbourne.
e. PYTHON.
These machines are representations of the Japanese Naval Attache machine. Six (6) have been completed; five (5) are in operation at the Naval Communications Annex and one (1) is on loan to the Army.
f. COPPERHEAD V.
COPPHERHEAD V is an electronic unit designed for extremely rapid differencing of cipher text. It was conceived as a fundamental machine in the breaking of subtractor ciphers. This machine has not progressed beyond the design and experimental stage.
10. In addition to analytical machines for use on the German and Japanese cryptographic systems, various projects for equipment of a general nature have also been carried on at Dayton. Among these are the following:
a. MIKE.
MIKE is a rapid digraph counter which takes its data from two (2) teletype tapes which run synchronously. It has proved extremely useful in the solution of the Hagelin ciphers.
b. RELAY CONTROLS AND COUNTING UNIT FOR THE GENERAL PURPOSE COMPARATOR.
These relay controls and counter units have been completed. Two (2) of these units are in operation at Naval Communications Annex, and the Army has been supplied with two (2) units.
c. GENERAL CRYPTOGRAPHIC WORK.
A special cryptographic project assigned by the Director of Naval Communications has been carried on at Dayton. Pilot models are now complete on this project.
d. ELECTRONIC ROTOR.
Approval has been granted for research on the problem of performing the functions of a commutator by purely electronic means using special tubes. Considerable thought has been spent in this direction but the research still remains in a fundamental state.
11. Curtailment of the contract for the additional Bombes has been accomplished. Cancellation of this contract resulted in the release of approximately two hundred (200) civilians and twenty-five (25) Waves from the Laboratory. These Waves have been returned to Washington. In estimating requirements after G-Day, the only feasible plans seems to be to maintain the Laboratory at its present level because of current projects assigned to it, either in connection with streamlining the Bombe equipment or developments in Japanese and general cryptographic fields. Present Naval staff at the Laboratory is as follows: two (2) Wave officers and fifty-two (52) enlisted Waves. The general personnel assigned to the Laboratory by ComNINE are as follows: 1 Ensign (Nurse), 1 Lt. (Med. Corps), 2 Lts. and 1 Lt. (j.g.)
Male officers, 2 Lt. (j.g.) Waves, 4 enlisted Waves (3 PhM, 1 Yeoman), 22 enlisted men. Should the load at Dayton increase or decrease, personnel may be sent to or from Washington on a temporary duty basis. It is true that the Laboratory facilities at Dayton may not be available in the post-war period. The only solution of the problem of keeping up-to-date with scientific developments and their applications to cryptanalytical problems is in maintaining a staff at the Naval Communications Annex who will develop projects to a point where they may be placed with private laboratories.
12. The following projects are planned for immediate assignment to the Laboratory at Dayton:
a. TWO ADDITIONAL DOUBLE UNITS.
Materials are available from the cancelled contract for the assembly of two (2) such units within two (2) months. As pointed out above, these double units have been particularly effective in connection with German U-Boats operating in the Far East. Recent knowledge of a German U-Boat operating in Australian waters came to light only through this equipment. Certain secret forms of the Berlin-Tokyo Naval Attache circuit can only be solved on this equipment.
b. PROVISION OF ADEQUATE MAINTENANCE STOCKS TO KEEP PRESENT BOMBE EQUIPMENT IN OPERATION FOR A PERIOD OF FIVE YEARS.
The provision of these maintenance stocks is felt justified by the threatened Japanese use of the Enigma as well as the use of the Enigma machine by other nations. New usage of the Enigma machine will probably necessitate certain modifications of the present equipment; materials and staff for these attachments should be available at the Naval Communications Annex.
c. ELECTRIC HAGELIN.
Present decipherment of Hagelin traffic in volume is extremely laborious using the general-issue Hagelin
machine. It is proposed to build an electric Hagelin which can operate on punched tape for rapid cryptanalytical analysis and exploitation of Hagelin material.
d. GENERAL PURPOSE COMPARATOR.
The experience with general purpose cryptanalytical machines developed under contracts with Eastman Kodak Company, Gray Manufacturing Company, and the I.B.M. Corporation, has been studied carefully. Although the present equipment has proved invaluable in the analysis and solution of certain cipher problems, some problems have shown the necessity for a machine of a more general nature. It has been attempted to incorporate in a single unit many of the important processing methods which present equipment has failed to accomplish.
e. SLIDE GRENADE FOR JAPANESE ENIGMA.
The captured model of the Japanese Enigma machine indicates that multi-notch wheels will be used. Present Grenade equipment must be modified to handle the problem on a production basis. It is planned to equip up to four (4) Bombes with the necessary attachments.
13. There are many cryptanalytical problems requiring further machine and development research, as follows:
a. STATISTICAL BOMBE.
Present Bombe equipment requires a crib. Theoretical studies have been made which indicate that it is definitely possible to construct equipment which will do all that a Bombe does but which will need no crib, the statistics of the language serving instead. Japanese practice of varied start positions in their messages makes placement of cribs difficult and the need for a statistical approach stronger. In any diplomatic usage of the machine the crib situation will probably be poor.
b. GEHEIMSCHREIBER.
This is a German cipher machine based on the
five (5) units of Baudot code. The British have constructed special machinery for the solution of this problem but their efforts have not been successful for the solution of the device. From various sources it has been inferred that the Germans were contemplating the Naval use of this device. As modern communication methods tend toward radio teletype, it is felt that encipherments similar to the GEHEIMSCHREIBER are of increasing importance and efforts should be made to standardize the solution.
e. SUBTRACTOR CIPHERS.
Rapid changes in Japanese cipher systems make a present appreciation of the situation difficult. Whatever direction these changes take, special equipment will probably be necessary. Theoretical studies are continuing in order to be in a position to take immediate action as soon as the Japanese policy becomes clear.
14. Reference (c) requested authority to pursue developments along electronic lines and was approved by the Director of Naval Communications. The fundamental points in the above discussion, upon which approval is assumed in accordance with reference (c), are as follows:
a. Maintenance of the present Naval staff at the U.S. Naval Computing Machine Laboratory, Dayton, Ohio, as outlined in paragraph 11 above.
b. Specific projects enumerated in paragraph 12 above.
Specific problems for future projects as outlined in paragraph 13 will be submitted when present theoretical studies are completed if the situation appears to demand action.
J.N. Wenger
Commander, U.S. Navy