The Man Who Broke Napoleon's Codes (30 page)

Marshal Marmont met his dinner guest that evening with some sense of curiosity and excitement. His servants were preparing his meal in unusual surroundings, a large house in Sabugal that had been requisitioned as French HQ. Marmont's incursion into northern Portugal had reached its high water mark and the supply situation being what it was, he would soon have to turn back into Spain. Still, he was not unhappy. He had followed the emperor's instructions while having the satisfaction of being proved right, since “laying waste northern Portugal” had not caused Wellington to adandon the siege of Badajoz. Secondly, he had a most interesting dinner companion, someone who might be useful in penetrating the miasma of uncertainty that enveloped his enemy's operations.

Marmont's guest was none other than Lieutenant Colonel Colquhoun Grant, Wellington's intelligence officer. To say that his presence at the marshal's table was reluctant would have been understating matters. For on 10 April, Grant had been captured. Sent north ahead of the main army to plot French dispositions, he had been caught between two patrols of French cavalry. Realizing they were surrounded, Grant and his Spanish guide, Leon, had dismounted and tried escape through a copse. This final gambit had failed and Leon, who was not wearing any uniform, had been executed on the spot.

As the cutlery clicked on the marshal's porcelain and crystal glasses full of wine were emptied, he had cross-examined Grant. Tell me more about the character of Wellington? Will he try to cut off my withdrawal into Spain? What is the spirit of your army? Marmont, it seemed, viewed his contest with Wellington as an intellectual challenge. He wanted to learn as much as he could about how the British general's mind worked. He wanted his guest to know, as he seemed to want everyone around him to know, that the English were doing battle with one of the finest minds of the French army. Grant met these enquiries with politeness, while concealing whatever knowledge he actually possessed of his master's plans.

As the meal drew to a close, Marmont evidently felt that his hunger for information had not been satisfied. He looked across at Grant and told him, “It is fortunate for you, sir, that you have that bit of red over your shoulders [meaning his uniform]; if you had not, I would have hung you on a gallows twenty feet high.” The marshal asked for, and obtained,
Grant's
parole,
his word, that he would not try to escape. Their interview was at an end and the captive was taken back to Salamanca.

While Marmont's army was menacing Almeida and generally driving the Portuguese about the countryside, militiamen and Don Julian Sanchez's guerrillas had swarmed about the French rear. As Wellington brought his HQ back to the north and Marmont began inching back into Spain, a packet of intercepted dispatches was delivered to the British commander. Two, dated from the Portuguese town of Sabugal on 12 April, shortly before he pulled out of that town, gave arrangements to Marmont's divisional commanders for the withdrawal back to the border. Fortunately for Wellington it was in an Army of Portugal
petit chiffre
and it is likely that Somerset was able to discover its details.

Wellington then acquired, in fairly quick succession, four messages in the higher command
Grand Chiffre,
the fourteen-hundred-character cipher circulated to the most senior officers a few months before. Two had come from Marshal Soult in Seville, and were taken from one of his ADCs. The first was long, with dense passages of the strong cipher, but from what they could make out it was evidently an essay of self-justification following the fall of Badajoz, since the fortress was located in his area of responsibility. Soult sought to blame Marmont for those events and displayed his own suspicions about the communications war being waged by Britain and its allies, telling Paris:

Soult's letter may be taken as evidence that the paranoid are not necessarily wrong. But he was not completely right either, since he did not consider that Wellington might already have resolved to finish the siege of Badajoz before intercepted letters from Marmont actually reached him. Marshal Soult had in any case demonstrated his confidence in the Great Cipher in writing his own letters. Shortly after his staff officer was captured, a further letter in this cipher, from Marmont to Berthier and
dated 16 April, was taken. Finally, a short message from General Dorsenne, commander of the Army of the North, to Marshal Jourdan in Madrid and also dated 16 April, was seized.

Scovell rode back into headquarters on 25 April after bidding fond farewells to Mary. She was going to stay in Lisbon until they could meet again. Like many an officer who ekes every last moment out of his leave, he was obliged to ride hard when he left the city and it took him just three days to travel from the Portuguese capital to Fuente Guinaldo, a few miles across the Spanish border, where Wellington's staff had once more taken up residence.

From late April through to early June, Wellington remained in the handsome mayor's house in Fuente Guinaldo's plaza. Upon returning north he had been disturbed to find that a Spanish garrison thrown into Ciudad Rodrigo had put down its tools because of arrears of pay, leaving January's breaches in the city's defenses unrepaired. Although he had taken possession of all of the main border fortresses by mid-April, Wellington did not feel free to make his next move until he was confident that Rodrigo would be able to resist a purposeful enemy attack. Once this work had been done, he would be free to launch himself into Spain.

Major Scovell, meanwhile, had returned to HQ and was able for the first time to consider several examples of the
Grand Chiffre;
Wellington had christened it the Great Paris Cipher. If he had expected Scovell to crack it with the celerity he had shown on Marmont's codes, then his lordship would certainly have been disappointed. Scovell looked upon the captured scraps of paper with wonderment. This was something altogether different from the Army of Portugal codes. A solution might be impossible. If this new cipher was changed regularly, as Marmont's had been, then it would certainly be. Can you produce the key or can you not, Major Scovell? He would have to study these messages most carefully before he could say.

None of the messages was entirely in code. That was one distinct advantage. What was not encoded often provided useful context, as he had learned with his earlier deciphering. Further study also allowed certain deductions to be drawn from the way certain passages had been enciphered—in particular, the longest and shortest of them. Soult's weighty catalogue of excuses of 14 April yielded something to the
reader, even if the meanings of its long ciphered sections were obscure. It had used hundreds of code numbers ranging from 2 to 1390. Many numbers were not used at all, others only used once or twice. Some, however, did repeat themselves rather a lot: 13 had been used twenty-five times; 210 was repeated thirteen times; 413 appeared in a dozen places and 2 in nine. These most frequent codes (13, 210, 413 and 2): Were they very common words, letters or bigrams? Could one of them be
et,
the most frequently used word in French?

When looking at how these frequently recurring numbers appeared in the other messages that fell into British hands in April and May 1812, some early hypotheses could have been formed. Marmont's message of 22 April, for example, contained the passage, “nothing will stop the Emperor's Army of Portugal 13.70.354 ….” In this context 13 might mean
de,
or
from
in this sense. The word of seems to come through quite clearly in a wish list (for various supplies) compiled by the duke of Dalmatia that featured the code 13 repeatedly. Elsewhere in Soult's message of 17 April, there is a passage of code ending “… 722.1074.13.821 campaign.” If 13 means
de,
then 821 must be
la
since the gender of the word campaign is feminine in French, and it finally becomes a fair working hypothesis that “13.821
campagne”
stands for
de la campagne.
A similar methodology, working across several dispatches, would have suggested that the commonly occuring 210 meant
e.

The shortest letter, Dorsenne's, was even more interesting. In its entirety it reads:

Dorsenne's dispatch was crassly enciphered, as bad in its way as the General Montbrun letter of August 1811 that had proven vital to the breaking of the Army of Portugal code. Given the much greater size of the
Grand Chiffre,
the damage resulting from this inept staff work was considerably less, but damage there was because this note could be used to begin a substantial attack on the Great Paris Cipher. Mixing in and out of code gives far too much away, an effect that is more pronounced in its original French, since the gender of nouns or form of verbs can reveal much.

Taking the second paragraph, which begins “In the letter of 16th March, 1207 announced 516.1264 was giving, … ” it is fairly clear that 1207 is a person. The construction “announced 516” is also interesting, since the normal use of language dictated the word
that
must follow. Does 516 mean
that,
or is it just the letter
t,
a bigram for
th
or even a blank, with 1264 meaning
that?

A little later on, we see “I would humbly assure 73.516.918 will not …” The code 516 is repeated. The sense of this second phrase is clearly something like: “I humbly assure you,” or HM or the King, whoever, “that …” There seems to be a likelihood 516 does indeed mean
that.
This becomes all the clearer with what follows this code number. In French, “918
ne negligerai aucune occasion,”
the tense and person of the verb
to neglect
makes it apparent that 918 must mean
je,
or L Examining the passage, it is clearly a good working hypothesis that 516 means
that
and 918 I.

As he sat struggling with this puzzle, Scovell's superb grasp of French grammar and syntax was his best weapon. His ability to absorb language was quite remarkable. Having landed in Iberia with Latin and some Italian, he had quickly obtained a sophisticated grasp of Spanish, noting in his journal following an evening's entertainment with Spanish hosts just a few months later: “I could easily perceive that the double entendre was the grand substitute for real wit.” With French, he had the benefit of having studied the language for twenty years and was therefore able to fill in the blanks in ciphered passages better than almost anyone else at headquarters could have done.

Going back to the first part of the second paragraph of Dorsenne's letter, if 516 on its own means
that,
then 1264 is likely to be another individual. Attacking these passages something begins to emerge: that numbers from 1201 up represent many individuals, names of particular armies and places.

It may be recalled that several months before, when the
Grand Chiffre
arrived at Joseph's headquarters, it had consisted of a 1750-type cipher table of twelve hundred numbers previously filled out in Paris. In expanding the table to include terms used in Peninsular warfare (by simply gluing on two farther columns of one hundred numbers each) the staff officers responsible had made the mistake of adding their new codes entirely in the columns marked 1201 to 1400. So 1201, the first of the coding numbers created by this expedient, stood for Malaga, 1202 for Valladolid, and various other numbers for the commanders and armies found in the Iberian theater of war. Certain military terms were also contained in this appendix of ciphering numbers: 1330, for example, meant
gunpowder;
1392 was the French National Guard.

It would have made more sense for Joseph's staff to have created a completely new cipher in which these terms were spread throughout the numbering as a whole, rather than being mostly confined to those between 1201 and 1400. Evidently, secret writing was a strange alchemy to the officers of Joseph's staff and they did not feel confident enough to alter the basic table that had been sent to them; it was evidently the fruit of endless cogitation by the savants at the foreign ministry. The effect of all this was that before Major Scovell or Wellington had ever clapped eyes on this Great Cipher, it had been weakened, albeit slightly, by the very people responsible for securing the French army's messages. It was not the case that
all
Peninsular terms and names were confined to that limited range of codes, because the original diplomatic code (sent out from Paris) did have entries for Madrid (it was 505) and a few other relevant places.