The fact that the sea-fire was made a State secret proves that it did not belong to the same family as the Greek fire of Æneas and Vegetius which, in one form or another, had been known all over the East from time immemorial. It was a new mixture—_i.e._ either a mixture containing some substance not hitherto known, or a mixture of known substances not hitherto combined together for warlike purposes. Many hold that an unknown substance was employed, and, further, that it was no other than saltpetre. We might, of course, fall back on the conclusion established in Chap. ii., and reply that saltpetre was not discovered until the thirteenth century and could not have been used as an ingredient of an incendiary in the seventh century. But the conclusion drawn in Chap. ii. was not a certain one: it was there characterised as highly probable. Saltpetre might possibly have been employed, and a belief which is shared in by some good writers deserves respectful consideration. We have, therefore, to investigate how far a saltpetre mixture would satisfy the above four conditions. There was absolutely nothing to attract public attention in the purchase from time to time of common, well-known substances, such as sulphur, quicklime, naphtha, &c. &c., by the authorities of the Arsenal; but the suspicions of the spies and traitors, always to be found in Constantinople,[68] would have been instantly roused by the importation of any new or rare substance such as saltpetre. And whence could saltpetre have come? M. Berthelot recognises the importance of this question, although he cannot answer it: “Comment se procurait-on le salpêtre?... Aucun renseignment n’est venu nous l’apprendre. Ce point pourtant est capital.”[69] Saltpetre would naturally have been obtained from the countries where it was most abundant and cheapest—from the East; but the Greeks could not have relied upon this source of supply, for whenever political complications arose between the Emperor and the Caliph—and they were interminable—the ports of Egypt and Syria were closed against Greek ships. However, saltpetre did not grow in the streets of Constantinople: the natural salt (if used) must have been collected _somewhere_, and sold to Government by _someone_, and transported _somehow_ to the capital; and what despot could have tied the tongues of collectors, merchants, sailors, and porters? The mere facts that only one State trafficked in saltpetre, that this State only bought it in time of war, and that this State alone employed sea-fire, would have immediately betrayed the secret of its composition to these men, and what was known to them was known to the world. It is most improbable that the use of saltpetre could have been concealed for one year, much less the five hundred years during which the secret of the sea-fire was successfully guarded. I may be reminded of the Emperor Constantine VII.’s statement (in Chap. xiii. of his “Administration, &c.”), that on one occasion a Roman