A Zoroastrian Educational Institute



HomeArticlesAuthorsBook ReviewCommunityLibraryProminentsRegisterStoreArticle SubmissionAbout Us




The Oil Fields and the Fire Temple of Baku


Jamshid Varza

Abraham Valentine Williams Jackson



Related Articles:

Related Links:


















From Constantinople to the home of Omar Khayyam; Chapter IV
Abraham Valentine Williams Jackson
The McMillan Company, 1911

 'By a towne called Backo, neere vnto which towne is a strange thing to behold for there issueth out of the ground a marueilous quantitie of Oyle.'
- JEFFBEY DUCKET, Fift Voyage into Persia, p. 439.

An oil spouter blowing off

A VISIT to the oil wells at the fields in the Balakhany-Sabunchy-Romany region, about eight miles northward from Baku, or at Bibi Eibat, some three miles south of the city, is an interesting experience. It was my pleasant privilege, during my first visit at Baku, to pay such a visit under the guidance of the. British vice-consul, Mr. Urquhart, whose kindness was equaled only by his knowledge of Baku and all that relates to the city.

The oil fields ablaze

On entering the fields, one becomes lost amid a maze of towering derricks, erected over the wells to operate them. These pyramidal wooden structures are covered with gypsolite or iron plating as a protection against fire. The shafts of the wells are often sunk fifteen hundred or two thousand feet to strike the oil. [1] Metal tubes, from six to twelve inches in diameter, are inserted in the bores so as to serve as pipes through which the precious liquid may spring upward or be drawn to the surface by a metallic bucket, the ' bailer,’ ready to pass in iron conduits to the refinery in Black Town and become one of the richest articles of commerce.

Inscription of the shrine of the fire temple, Baku

There are over two thousand of these wells in the Apsheron (Absharan) Peninsula, on which Baku stands, and one can hardly conceive of the activity implied in this wilderness of truncated pyramids, each in itself a source of revenue that is a fortune.[2] The stupendous figures of the annual yield from these fields is almost staggering. It runs up into the many millions of 'poods,' a pood being approximately five American‘ gallons.[3] Sometimes the borings strike ‘fountains,’ and then a tremendous 'spouter’ is the result, belching up its concealed contents with the force of a geyser, and perhaps bringing ruin instead o f fortune to its owner unless the giant can be speedily throttled and gagged. Thrilling descriptions are given of how some of these monsters have, within the last generation, thus wrought destruction to everything within immediate reach.[4] The magnificence of the spectacle is surpassed only by the awful grandeur when fire adds terror to the scene. On all occasions when visiting the petroleum fields it is advisable to wear old clothes, for one may find it anything but the oil of gladness, as I learned from sad experience when one of the spouters blew off unexpectedly while we were near it, filling the air with a deluging rain from whose greasy downpour there was no escape.

The fire temple and its precinct at Surakhany, Baku

All who have been at Baku, and almost all who have heard of Baku, know of the Fire Temple at Surakhany in the northern environs of the city. The place is easily reached by driving, or better by rail with a short spin in a phaeton afterwards. Common tradition has long associated the one-time sanctity of this region with the veneration of the so-called Zoroastrian Fire-worshipers, though whether with absolute justice must remain to be seen.

Zoroaster flourished at least as early as 600 B.C., and his religion became the faith of Ancient Iran, continuing as its creed until the seventh century of the Christian era. This remarkably pure religion, which bears striking resemblances to Judaism and to our own faith, which it antedates, was supplanted in the seventh century of our era by the new and militant creed of Islam. Ormazd yielded his throne in heaven to Allah, the Avesta gave place to the Kuran, and Zoroaster was superseded by Muhammad as the acknowledged prophet of truth. Having previously devoted a volume to the life and legend of Zoroaster and a monograph to his religion, as well as given special attention to the subject as a whole in my former book on Persia, I shall not now go further into the history and fortunes of the ancient creed.[5] Suffice it to say that a few Zoroastrians refused to adopt Muhammadanism when the conquest came (650 A.D.). Some of this scanty band sought refuge and freedom to worship Ormazd in India, where they became the ancestors of the flourishing community of the Parsis in Bombay; a remnant persisted in staying in their old home, only to meet with persecution and hatred as infidel Gabrs, 'Ghebers, Unbelievers,’ and these still find an insecure asylum at Yazd and Kerman in the desert, while a handful even reside in Teheran.[6] They see not God but the purest effulgence of God in the Flame Divine, and they abhor the name of 'Fire-worshipers.’ Yet in the eyes of Mohammedans they are such, and it is not strange that local tradition associates their name with Baku as the very source of eternal fire.

Inscriptions XII and XIII on the walls of the precinct

I have said that ' tradition ' has connected the name of the Zoroastrians with the igneous realm of Baku, but I have not been able to trace it back more than two hundred years, as I shall show below, and I believe that some of the sweeping statements made on the subject by modern writers (including myself) may have to be modified so far as Zoroastrianism is concerned. The present shrine is apparently of Northern Indian rather than of Persian foundation, although possibly the site itself may have been a hallowed one in ancient times; but before I turn to that matter, I shall give a description of the sanctuary and its surroundings.

The sacred precinct consists of a walled enclosure that forms nearly a parallelogram, following the points of the compass. Its length is about thirty-four yards from north to south, or forty on its longer side; the breadth is about twenty-eight yards from east to west. The central shrine stands nearly in the middle of the court. A square-towered building, approached by a high flight of steps, rises toward the northeast corner. The walls of the precinct are very thick, as they consist of separate cells or cloistered chambers, running all the way around, and entered by arched doors. The whole is solidly built and covered with plaster.

The central sanctuary, eastern elevation, with inscription.

The structure in the middle is a square fabric of brick, stone, and mortar, about twenty-five feet in height, twenty feet in length, and the same in width, with arched entrances on each side facing the points of the compass.[7] These entrances are approached by three steps each on the north and east sides, and by two steps on the south and west sides, where the ground is slightly higher. In the middle of the floor is a square well or hole (visible in my smaller photograph), measuring exactly forty and one-half inches (1 m. 13 cm.) in each direction. Evidences are seen of pipes once used to conduct the naphtha to this and to the roof. The top of the shrine is surmounted by four chimneys at the corners, from which the flaming gases used to rise when the temple was illumined in times gone by. In the middle of the roof is a square cupola, from whose eastern side there projects, like a flag, a three-pronged fork that resembles the trisula, or trident, of the Indian god Siva.

High over the archway on the eastward front is a double oblong tablet, three and a half feet high by two broad, the upper section of which shows a swastika emblem and a sun, four flowers, and several nondescript figures.[8] The lower section is devoted to an inscription in nine lines in the Nagari character of India, beginning Sati Sri Ganesaya namah, 'In verity, Homage to the Honored Ganesa ‘the common invocation, in Sanskrit writings, to the divinity who removes obstacles. The inscription continues, apparently in the Marwar dialect of the Panjab,[9] stating that the shrine was built for Jvalaji (the same as the flame-faced goddess Jvala-mukhi, of Kangra in the Panjab),[10] and quoting a Sanskrit couplet on the merits of a pilgrimage and pious works. It concludes with the date of the Vikramaditya era, 'Samvat 1873‘ (=1816 A.D.).[11] As the inscription is high and in a position not easy for photographing, it was necessary to perform the somewhat acrobatic feat of perching on the top of a long ladder that was shakily held in the air, about ten feet off, by six men. Considering the difficulties of the task, the photograph (the only legible one I have seen) turned out to be quite a success.

The portal and tower edifice in the eastern wall, fire temple of Baku

Around the walls of the precinct, either above or near the side of the doorways of the cells, are fifteen more dedicatory tablets sunk in the plaster and written, with one exception, in the same Nagari character, prevailingly used for Sanskrit, or in a variety of this Indian alphabet. In my notebook I gave them numbers, beginning at the northwest corner near the usual entrance. Two of them in the northern wall are in the Panjabi language and script of Upper India and are of Sikh origin, as they quote from the Adi Granth, the sacred book of Nanak’s religion, which was founded about 1500 A.D. Their date, however, must be two centuries or more after that era, as was shown by Dr. Justin E. Abbott, of Bombay, from a couple of photographs of two other tablets in the southern wall, which I brought back after my first visit to the temple.[12] One of these latter inscriptions (XII = also published previously by Colonel C. E. Stewart in JRAS. 1897, pp. 311-318) is in Nagari and bears the date 'Samvat 1802' = 1745 A.D.; the other (XIII), immediately below it, is in Persian, the only one in that language, and gives the same year according to Mohammedan reckoning, Hijra '1158' = 1745 A.D.[13] I have since found similar dates on still others of the tablets within the precinct. For example, one of the tablets (I) near the west-by-north corner, beginning Sri Rama sat, contains the year 'Samvat 1770' = 1713 A.D., and seems to be the earliest date found. One of the northern tablets (IV) appears to have the date 'Samvat 1782' (?) = 1725 A.D., although the figures for 8 and 2 are very uncertain. Another (XI), on the east side, bearing a swastika followed by Om Sri Ganesaya namah, has at the end (though almost illegible)' Sam(v)at 1820' = 1763, Yet another (IX), also on the eastern wall, concludes, 'Samvat 1839' = 1782 A.D. (the last cipher being slightly broken); while still another (V), on the northern side, closes, as I read it, with '1840' = 1783 A.D.[14] They all belong, therefore, to the eighteenth century.

Worshipers in the Baku temple in 1805

Besides the total of fifteen inscriptions, dated or undated, within the circumvallation, there are two more on the outside, connected with the square tower near the northeast corner. The lower one of these (XVI) is inscribed on a black stone over the arch, facing east ; it begins with the common Ganesh formula, mentions Jvala-Ji, the divinity of fire, and concludes with the date 'Samvat 1866’ = 1809 A.D. The upper one (XVII), placed on the outside of the story above, is likewise in Nagari, but is less clearly written, and the close is hardly legible in my photograph, which was taken with difficulty, as I had to be held up over the top of the high doorway while I made the snapshot. The tablet has been stupidly set in, or reset, upside down, and below it is scrawled in Russian block letters the name 'N. Mintova.’

As I secured photographs of all the inscriptions, and the majority were successful, in case the tablet photographed was still legible,[15] it will be possible to study them in greater detail later, and I hope to publish them elsewhere in cooperation with Dr. Abbott, whose interest in the subject has already been proved.[16]

After this long, and, I fear, somewhat technical, disquisition, I turn to the more attractive problem of determining the possible age of the temple and its buildings. I may state at once that I used to hold the generally current opinion that the sanctuary was of Gabr, or Parsi, origin -a Zoroastrian fire-temple. [17] Further study of the subject has forced me to abandon this view (certainly for the present temple) as the following paragraphs will show.

So far as my researches go, I have not been able to find any allusion to a temple on the site in the classic writers of Greece and Rome; [18] nor in the early Armenian authors;[19]3 nor do the medieval Arab-Persian geographers refer to it, as we might expect, when mentioning the naphtha wells ; nor again is it spoken of by the European travelers Barbaro, Jenkinson, or Ducket in the sixteenth century, nor by Olearius in the seventeenth century, nor by John Bell early in the eighteenth, when touching on Baku, nor yet, earliest of all, by Marco Polo -- all of whom have been cited above. The oldest reference I have been able thus far to discover (though I stand ready for correction) is by Jonas Hanway in the year 1747, who was practically contemporaneous with the inscriptions given above. Hanway himself did not visit the temple when he was in Baku, but he gives a detailed and accurate account of it, based upon 'the current testimony of many who did see it.’ He speaks of the worshipers as ‘Indians, 'Gaurs,' or 'Gebrs,’ and devotes a chapter to describing the religion of Zoroaster, somewhat in detail.[20]1 The heading of this chapter reads: --

‘A succinct account of the antient PERSIAN religion, with several minute particulars relating to the everlasting fire near BAKU, and the extraordinary effects of this phsnomenon, to which the INDIANS pay divine honours.’

Scene at fire temple of Baku, 1825

After giving some idea of Zoroaster and his doctrines as followed by the early Persians, he adds: - 'These opinions, with a few alterations, are still maintained by some of the posterity of the antient INDIANS and PERSIANS, who are called GEBERS, or GAURS, and are very zealous in preserving the religion of their ancestors; particularly in regard to their veneration for the element of fire. What they commonly call the EVERLASTING FIRE, near BAKU, before which these people offer their supplications, is a phaenomenon of a very extraordinary nature, in some measure peculiar to this country, and therefore deserves description. This object of devotion to the GEBERS, lies about 10 ENGLISH miles north-east by east from the city of BAKU, on dry rocky land. There are several antient temples built with stone, supposed to have been all dedicated to fire; most of them are arched vaults, not above 10 to 15 feet high. Amongst others there is a little temple, in which the INDIANS now worship: near the altar, about 3 feet high, is a large hollow cane, from the end of which issues a blue flame, in colour and gentleness not unlike a lamp that burns with spirits, but seemingly more pure. These INDIANS affirm, that this flame has continued ever since the flood, and they believe it will last to the end of the world; that if it was resisted or suppressed in that place, it would rise in some other. Here are generally forty or fifty of these poor devotees, who come on a pilgrimage from their own country, and subsist upon wild sallary, and a kind of JERUSALEM artichokes, which are very good food, with other herbs and roots, found a little to the northward. Their business is to make expiation, not for their own sins only, but for those of others, and they continue the longer time, in proportion to the number of persons for whom they have engaged to pray. They mark their foreheads with saffron, and have a great veneration for a red cow. They wear very little cloathing, and those who are of the most distinguished piety, put one of their arms upon their head, or some other part of the body, in a fixed position, and keep it unalterably in that attitude. A little way from the temple is a low clift of a rock, in which there is a horizontal gap, 2 feet from the ground, near 6 long and about 3 feet broad, out of which issues a constant flame, of the colour and nature I have [p. 382] already described : when the wind blows, it rises sometimes 8 feet high, but much lower in still weather : they do not perceive that the flame makes any impression on the rock. This also the INDIANS worship, and say it cannot be resisted but it will rise in some other place. About 20 yards on the back of this clift is a well cut in a rock 12 or 14 fathom deep, with exceeding good water.’ [21]

The descriptive portion of this account, as already stated, is correct, being based upon information received from accurate observers. It is plain from the description itself that, if actual Gabrs (i.e. Zoroastrians, or Parsis) were among the number of the worshipers at the shrine, they must have kept in the background, crowded out by Hindus, because the typical features which Hanway mentions are distinctly Indian, not Zoroastrian. The allusion to the tilak mark on the forehead, the veneration paid to the red cow, the diet of herbs and fruits, the scantiness of clothing, and the Yogi posture of the urdhva-bahu ascetics, with withered arms, are all Brahmanical.

Further external evidence of the same character may be gained from the testimony of succeeding travelers. Thus, S. G. Gmelin (1771) describes the various Yogi practices of the devotees, especially of one ascetic who had held his arm up for seven years, until it became stiffened -a species of self-castigation that is only Hindu and was never sanctioned by Zoroastrianism. [22] Similarly, Jacob Reineggs, who made several journeys in the Caucasus before 1796, in describing the 'Ateschjah,’ speaks of the devotees as 'Indianer,’ who were formerly called 'Geber '; and he mentions their Yogi austerities, noting also that they burned their dead -- a fact sufficient in itself to prove they could not have been true Zoroastrians.[23] Pointing in a like direction is an incidental reference by Morier (between 1800 and 1816), when he casually mentions meeting a ' Hindu pilgrim ' returning from Baku to Benares.[24]

The scholar Eichwald (1825-1826), who gives a clear description of the temple and some of the ceremonies, mentions the names of Hindu divinities as invoked in the worship – Rama ('Rahma'), Krishna ('Krisehni'), Hunuman ('Hanuma'), and Agni (' Aghan '), the god of fire -- all of which are Brahmanical, as is likewise the blowing on the conch shell ('Tritonmuschel') in the ritual. His picture of the temple, with its naked worshipers (here reproduced), his reference to the Kangra temple in India, and above all, his mention of a place in the cloister where the devotees burn the body of any of their number that may die, leave no doubt as to the Hindu character of the shrine at that day. He himself (pp. 216-217) properly emphasizes the fact that the Indian fire-worshipers at the temple had wrongly been called Gabrs ('Gueber').[25]

The German poet and student Friedrich Bodenstedt, writing in 1847, after spending seven years in Russia and the Caucasus, in telling of the 'Ateschgah,’ speaks of the Hindu god Vishnu ('Wischnu’) and of the ritual summons with the 'Tritonmuschel,’ and looks upon the idolatrous worship and barbarous self-mortification of the body as if it were a decadence from the exalted religion of Zoroaster (‘die erhabene Lehre Zerduscht’s'), whereas it was really only Indian.[26]

In November, 1858, the noted French writer Alexandre Dumas visited the temple. Throughout his description he assumes that the sanctuary was a fane of the Zoroastrian fire-worshipers ; and he refers to its ministrants as ' Parsis,’ 'Guebres,’ and 'Madjous,’ or descendants of the Magi. But it is clear from his description of the ritual, which he himself calls 'une messe hindoue,’ together with his allusion to the frequent recurrence of the divine name ‘Brahma' in the chant, the employment of cymbals, and the use of prostrations in the service, that the worship was simply Hindu.[27]

Petzholdt (1863-1864) gives an almost equally detailed description of the sanctuary and the ' Hocus-pocus’ ceremonies that were performed, but has nothing to show that there was anything Zoroastrian in their nature.[28]

About the same time as Petzholdt, the Englishman Ussher visited the temple, on Sept. 19, 1863 (or 1864 ?), calling it 'Atesh Dja,’ the Arabic pronunciation ’ of the Parsi name Atash Gah.[29]4 His diary notes were not originally prepared for publication, and he disclaims for his volume any pretence to special scientific acumen; but while he supposes that the pilgrims to the temple were ‘devotees from among the fire worshipers of Persia and India' (p. 207), and even though he mentions Zoroaster and the god Ormazd, his testimony, like that of his predecessors, to the effect that the steps of the little altar in the priest’s cell were ‘covered with brass and bronze images' (p. 208), bears on the face of it the evidence that the ministrant was a Hindu. His own words, in fact, state this when he appends : 'The present inhabitants were only two in number, both from India, one being a native of Calcutta, the other of Delhi… They wore the usual Indian dress and turban, having in addition a streak of yellow paint on the forehead between the eyes.’These priests are – pictured in the colored frontispiece to his book as engaged in performing their ritual on the pedestal of stone and mortar which is near the central shrine, and which, like it, is represented as lighted up with natural gas; their type is thoroughly Indian, and the ceremonies which are described are Brahmanical, not Zoroastrian.

Ten years later than Ussher, the German Baron Thielmann visited the fire-temple in October, 1872.[30] He calls it by the same name, 'Ateschgah,’ and says (p. IO): ' The priest is sent here for a limited time by the Parsee community of Bombay;[31] after a lapse of some years he is replaced. Now and then a pilgrim from the end of Persia (Yazd, Kerman) or from India makes his appearance and remains for several months or years at the sacred place.’ Proceeding on this assumption, the baron supposed that the worship was that offered to Ahura Mazda, or Ormazd; but it is manifest that the ritual which he witnessed and briefly described, specially the finale of ' a votive offering of sugar-candy made to an idol on the altar,’ was wholly Hindu, never Parsi. Thielmann’s further testimony (gathered through an interpreter), to the effect that ' the priest, according to his own statement, was ignorant of Zend and also of Sanscrit ' (p. 311), would militate against the celebrant’s having been a Zoroastrian dastur, who would surely have known the Avesta. As regards the statement that ' he understood Hindoostani, Hindi, and presumably also Parsee,' we may readily believe the first part, though the assumption as to the Parsi dialect is far less probable; and we may well agree with Stewart (p. 314, cited below) that Thielmann was mistaken.[32]

Still more strong in favor of the Hindu, not Zoroastrian, character of the present sanctuary is the internal evidence of the inscriptions which have been mentioned already. The first to draw my attention to this fact was the Parsi priest Shams Ul-Ulama Jivanji Jamshedji Modi, of Bombay. In a letter written to me in 1904, after my first visit, and from which I quoted later in print (JAOS. 25. 304), he expressed his doubts, from the Zoroastrian standpoint. These doubts (which had been anticipated long before by Eichwald -- see above, p. SO) were further strengthened by Dr. Abbott’s reading of three of the inscriptions, as already mentioned ; and they are now wholly substantiated by the other inscriptions, here made accessible. They are all Indian, with the single exception of one written in Persian (see my reproduction), which is dated in the same year as the Hindu tablet over it, as explained above. The Iranian tablet is a quatrain in not very good Persian, the mistakes of which might have been made by a Hindu imperfectly acquainted with the language, although Persian is current in northern India.[33]

Not only this evidence, but also the theory of the Indian origin of the shrine at Baku, was anticipated over a decade ago by Colonel Stewart in an interesting article entitled an 'Account of the Hindu Fire Temple at Baku,’ which was published, with a reproduction of three of the tablets, in 1897. Stewart had visited the place in 1866 and again in 1881, and he adds convincing evidence to show that the sanctuary is Northern Indian in source. He speaks of 'Hindu visitors who came here after visiting the Temple of Jawala Mukhi in the Kangra District of the Punjab. The Kangra Temple of the Flame-faced Goddess is well known in India.’[34] He further states (pp. 311-312) that, when he saw it 'in 1866, one Hindu priest alone watched the fire, although previously three Hindu priests had always watched.’ One of these had been murdered by the Muhammadans for his pittance of money; the other had fled. The third, who remained, spoke Panjabi, the language of at least two of the inscriptions, and had piously served at Surakhany for many years as priest of 'this greater Jawala Ji,’ as he called the divinity of flame, a name that appears several times in the inscriptions mentioned above, and that of course is connected with the Indian temple at Kangra. Colonel Stewart adds that when he returned to Baku in 1881 and again visited the temple, he found the fire extinguished and no priest in attendance. He furthermore states (p. 314) that near the Afghan border he met two Hindu Fakirs who announced themselves as 'on a pilgrimage to this Baku Jawala Ji’; and that in 1882, when he was returning to England, some of the Hindu traders begged to be allowed to accompany him as far as Baku for the purpose of visiting the shrine. He supplements this by saying, 'although the Hindus I have met in Persia know about this temple, I never heard any Zoroastrian in Persia, although I met many, express any wish to visit it or have any knowledge of its existence.’ His conclusion on this point is (p. 313), ' there can be no doubt this temple is not, or never can have been a Zoroastrian temple.’ The structure in the middle of the sacred enclosure he regards (p. 312, cf. p. 315) as 'a much more modern building,’ and he considers it to have been‘dedicated to the God Siva, as shown by Siva’s iron trident, which was fastened on the roof.’[35] It was, however, dedicated to some form of the Hindu divinity of fire, as we have seen.

Additional weight from the Hindu side is given by the style of architecture of the building. Judged from this standpoint also, irrespective of anything else, the structure appears to be Indian rather than Persian. Heinrich Brugsch, who spent part of a day at the sanctuary, about 1884, but says little about it, noticed that the edifice was built in Indian style (' in seinem indischen Baustyl'),[36] and it gives an impression similar to that of a Hindu Dharmasala, or religious building founded as an act of piety or charity. Stewart accordingly speaks several times of the whole sanctuary as a 'Dharamsala ' (p. 313), and remarks that it did not resemble the remains of any real Zoroastrian temple he had seen in Persia. I think we may readily share his opinion, even if the architecture of the Persian chapar-khaneh, or caravansarai, may possibly have exercised some slight influence on the style of the enclosure, and even though one might be inclined to see remote affinities with the ruined shrine near Isfahan and that near Abarkuh.[37]

From all this I believe that, even against our will, we must reach the conclusion that, whatever the site may possibly have been originally, the Baku fire-temple, as we now have it, is a Hindu product, and that it is, more particularly, of Northern Indian origin, where fire-worship was cultivated from the ancient time of the Vedas. In age the sanctuary can hardly be more than two centuries old, if we may judge from the half dozen inscriptions that are dated, as they belong mostly to the eighteenth century.[38] We may account for its presence at Baku more easily from the fact that ' formerly many merchants lived here, especially Indians,’ as is stated by Hanway (or rather by Cooke’s diary) in 174’7 ; and Eichwald (1825- 826) states that in his time the special patron of the temple was a rich Hindu, named 'Otumdshen’ (perhaps Skt. Atmajanma), who farmed the Caspian fisheries and lived mostly at Astrakhan.[39] I have already (p. 31) mentioned the fact that caravans from India were common in the region from early times.

Thus, to our regret,  vanishes the legend of the 'Zoroastrian' Atashgah at Baku, at least in the form in which we have it. The sacred flame that was its source has likewise vanished, for in 1879 the temple passed over into Russian hands by a concession of the government, when the last priest sold out his interests to the Baku Oil Company near the old Kokorev refinery, and the fire was extinguished forever.[40] It is true that photographers may still have the fane illuminated on occasions so as to give a more characteristic picture, or may paint into their plates the blazing naphtha for more realistic effect; but even in 1381, when Stewart last visited the shrine, it was under lock and key as now, and the engineer in charge of the Russian refinery ' relit ' the fire, only to extinguish it when he left the building, because, as he said, 'he wanted all the natural petroleum gas for heating the furnaces of his works ' (p. 312). Sic transit gloria ignis! -the flame has perished, a victim to the commercial value of the precious substance that gave it birth.

And one thing more. By a strange chance, the bell which once hung aloft from a hook that is still visible in the roof, and which was rung to mark the progress of the pagan ritual, now swings in the belfry of a Russian church which was in need of a signal to summon its worshipers to Sunday service.

[1] A good idea of the difficulties connected with making these borings may be gained from Norman, All the Russias, pp. 219-227, New York, 1904. 

[2] See Henry, Baku, pp. vii, 104, and the remarkable statistics given throughout that standard work.

[3] Through the kindness of Mr. Wilson D. Youmans, of Yonkers, N. Y., I have been most courteously furnished, by the statistician of the Standard Oil Company of New York, with the following figures regarding Baku : 'During the year 1908 the production of the Baku Field was 465,343,000 poods, equaling 55,863,504 barrels of 42 gallons. The world’s production of crude oil during the year 1908 was 284614,022 barrels of 42 gallons, and the Baku production represented 19.6% of the total production of the world. 'The production of the Baku Field annually since 1903 was as follows : -

1903  .  . . 696,581,155 poods

1906 . . . . 447,520,000 poods

1904 . . . . 614,115,445 poods

1907 . . . . 476,002,000 poods

1905 . . . . 414,762,000 poods

1908 . . . . 466,343,000 poods

‘During the year 1908 the production of American crude oil was 179,572,479 barrels of 42 gallons. The United States government uses as a basis 8.33 poods of crude oil = one American barrel of 42 gallons. This is approximately 6 American gallons to 1 pood.

[4] See Henry, Baku, pp. 63-108.

[5] See Jackson, Zoroaster, the Prophet of Ancient Iran, New York, 1899; Die iranische Religion, in Grundr. iran Philol. 2. 612-708, Strassburg, Persia Past and Present, New York, 1906.

[6] See Jackson, Persia, pp. 363-M.

[7] My measurements on the occasion of my last visit gave in meters, 6 m. 10 cm. x 6 m. 6 cm. or 20 ft. 0 in. x 19 ft. 11 in.

[8] These figures may represent a fire-altar, supposed to be behind a trisula, and possibly a chaitya-bell over a trisula. This is the view of Dr. Gray, Dr. Abbott, and myself. The exact measurements of the tablets in meters are: 1 m. 7 cm. high by 59 cm. broad, the upper section being 37 cm. in height, the lower 76 cm. in height. 

[9] Such is the opinion of Pandit D. Kosambi, of Poona, India, to whom, as to Professor Lanman, Dr. Gray, and especially Dr. J. E. Abbott, I am indebted for suggestions and help in connection with the decipherment. 

[10] See Stewart (and Cust), The Hindu Fire Temple at Baku, in Journ. Roy. As. Soc. 1897, p. 311-318. Several references to jvalamukhi in Sanskrit literature will be found in Boht-lingk and Roth, Skt. Wb. 3. 171-172. 

[11] There is some uncertainty in regard to the third cipher, whether 1 or 7. I have since found in Eichwald (Reise auf dem Caspischen Meere, 1. 217, Stuttgart, 1834) that in 1825 (or 1826) he saw and talked with the priest who, he says, had composed the inscription. So far as Eichwald could understand, the name of this recluse was ' Atteit Kanzenger,’ from the city of ' Kotessur ' (Hot Isa Shah, [Isvara ?] in the Panjab, N. India) ; that the temple had been erected '16' years previously, and the inscription on it began ' Ssri Gnass ( (i.e. Sri Ganesaya), mentioning the name of the Indian ruler ' Bikker Mandit ' (i.e. Vikramaditaya) and likewise his own name. I cannot find the priest’s name on the tablet, but Eichwald acknowledges the difficulty in understanding what he said. 

[12] Abbott, Indian Inscriptions on the Fire Temple at Baku, in JAOS. 29. 299-304, New Haven, 1908.

[13] For a rendering of the Persian tablet see below, p. 63, n. 2.

[14] Of this latter tablet I have two photographs.

[15] Inscription X (small) on the east side was wholly illegible, but I photographed it also. As most of the tablets are of cement, and the letters raised, they are peculiarly vulnerable to the attacks of the elements.

[16] I did not know until after my last visit to Baku that copies of the inscriptions had been drawn in 1860-1861 by Dorn, Atlas einer Reise im Kaukasus, St. Petersburg, 1871, reissued in 1896 by the press of the Royal Academy of St. Petersburg, in Russian, and also in German as Atlas zu Bemerk ungen einer Reise in dem Kaucasus und den sudlichen Kustenlandern des Kaspischen Meeres in den Jahren 1860-1861, St. Petersburg, 1996. I have worked throughout from photographs and personal inspection of the tablets, having secured Dorn's work only after this chapter was set up.

[17] See Jackson, Notes Journey to Persia, I, in JAOS. 25. 176-178.

[18] The Byzantine writer Priscus (d. after 471 A.D.) incidentally refers to the region of the naphtha wells of Baku, but not the temple, in an allusion to 'the flame that comes out of the rock beneath the sea' (Greek Text) in Frag. 8 (ed. C. FHU. 4. 90 b, and ed. Dindorf, 1. 313). The statement by Henry, Baku, 1.26, who cites Gibbon (ch. 46) as his authority, to the effect that the Roman emperor Heraclius appears to have destroyed the temple at Baku along with others, is due to a misconception. The temple which Heraclius destroyed was at Shiz (now Takht-i Sulaiman), and was not the sanctuary at Baku ; see Jackson, Persia, pp. 141-142.

[19] One might be inclined to see a reference to Baku in the allusion by the Armenian writer, Moses Rhorene (6th century or later), to a fire altar in Bagavan that was fostered by the Sasanian king Ardashir. The reference (Mos. Khor. 2; 77) reads in the German translation by Lauer, Geschichte Gross-Armeniens, p. 136, Re-gensburg, 1869, 'Auch vermehrt er [d. h. Artaschir] noch den Tempel-dienst und befiehlt, das Feuer des Ormist, welches auf dem Altar in Bagavan war, ohne erloschen leuchten zu lassen.' The modern Armenian author Alishan (Hin Havatk gam Hetanosagan gronk Hayots [i.e. Ancient Beliefs, or Pagan Religions of Armenia], pp. 5051, 302, 432, Venice, 1895) understands the passage as meaning Baku ; but both the German authorities, Hubschmann (Altarmenische Ortsnamen, in Indogermanische For-schungen, 16. 411) and Marquart (Eransahr, p. 110, n. 2), seem to allow of no question that Bagavan (Bagvan) is a small village, now Turkish Uich-Kilise, near Diadin and Mount Npat

(Ala-dagh), north-north-east of
Lake Van, a hundred miles from Baku. So that idea (otherwise attractive) may be dismissed as misleading.

[20] Hanway, Caspian Sea, bk. 3, ch. 67, 1. 379-384 = 3 ed. 1. 261-266. I ’ repeat again that in general ‘the description of the temple precinct is accurate, as based on trustworthy in-formation. Not all that is said about Zoroaster, however, would be accepted today. As already stated (p. 32, n. 1, above), Dr. John Cooke,  Travels, 2. 382, impugned Hanway's statements, on the ground of misrepresentation; he declares that he will not abuse his readers’ patience by an account of Zoroaster and his successors, nor imitate Hanway in the description which he had given of different vaulted temples 'from 10 to 16 feet high, which do not now exist and which probably never existed ' [this latter criticism. Is. unjust, as Hanway simply refers to the. cells] ; nor will he describe the horizontal gap in the rock from which a flame issues; but he will confine himself to the facts of his visit to the celebrated fire. I cite his remarks from the footnote of the French translation of George Forster, Voyage du Bengale a St. Petersbourg, tr. L. Langles, 2. 357-368, Paris, 1802, as Cooke’s English original is not accessible to me. The French runs: 'Le 11 fevrier 1747, notre ambassadeur, avec plusieum personnes de sa suite, alla voir ce feu celebre. Apres etre descendus des montagnes dans la plaine situee au nord de ces memes montagnes, a cinq ou six verstes au plus de Bakou, ilsentrerent dans un petit fosse carre b&i en pierres, dont l'aire avoit a peine un demi0acre, meaure d'Ecosse. Le sol Btoit compose d'un sable pur et leger; dans cette enceinte Btoit un puits. La surface de l'eau qui sortoit de cette source etoit couverte de naphthe blanc ; mais quelques pouces plus bas que la surface generale de l'aire du sable. Notre compagnie ne vit qu'une seule miserable [1.3581 salle, oh l'on dit que ces merveilleux edifices se voient, et un autre appartement fort mesquin, oh demeurent les religieux. Ils n'etoient pas alors plus de 40. Ils introduisirent tres-volontiers nos mes-sieurs dans la chambre dont je viens de parler. Il y avoit une place &pa-r&, comme nos choeurs dans les temples protestans, et que vous pouvez nommer leur autel, comme fait Hanway: quelques cannes creuses Btoient nlantees dans ce sable pur. Une de ces Cannes, plus grande et plus grosse que les autres, etoit placee au milieu, par maniere de preeminence. D’autres cannes Btoient tres-pressees autour de cellela, de maniere a former trios ouvertures au sommet, pour livrer passage a trois flammes bleues-pales. A l'epoque dont je parle, plusieurs de ces cannes etoient eteintes. Mais afin que l'ambassadeur et sa compaguie pussent voir l'effet que ces cannes pro-duisoient quand elles Btoient en activite, on apporta un vase rempli de naphthe pur; on en mit un peu sur le sable a l'entour des roseaux, et par le moyen d'un morceau de papier allume, le naphthe exhala de la flamme a travers les roseaux.'

[21] Hanway, Sea, 1. 381-332 = 3 ed. 1. 261-265.

[22] The reference to S. G. Gmelin, Reise, 3. 46, is taken from Eichwald, Reise auf dem Caspischen Meere, 1. 178-179, n., Stuttgart, 1834, as I have not been able to obtain Gmelin's own work.

[23] See Reineggs, Allgemeine Beschreibung des Kaukasus, ed. F. E. Schroder, 1.169, Gotha, 1796.

[24] Morier, Second Journey through Persia, 2. 243, London, 1818.

[25] See Eichwald, Reise auf den Caspischen Meere und in den Caucasus, 1. 176-183, 189, 216-217, Stuttgart, 1834. The only posaihle Zoroastrian feature in the whole description is the employment of the old name 'Ateschgah' (fire-temple) and the allusion (p. 181) to the hatred manifested towards mice, frogs, lizards, and snakes, 'als Kinder des bosen Geistes,' although this information may have been derived from some other source, and applied in this connection.

[26] See Bodenstedt, Volker des Kaukasus, pp. 137-139, 2 ed. Frankfort on Main, 1849 (preface dated Lago di Como, 1 Nov. 1847).

[27] Dumas (pere), Impressions de voyage: le Caucase, 2. 25-30, Paris, 1880.

[28] Petzholdt, Der Kaukasus. Studien in den Jahren 1863 und 1846,1. 207-210, Leipzig, 1866.

[29] Ussher, A Journey from London to Persepolis, pp. 206-207, London, 1866. It may be added that a not very accurate small woodcut of the temple is found in Bicknell's Hafiz, p. 337, London, 1875.

[30] Thielmann, Journey in the Caucasus, Eng. tr. by Heneage, 2. 9-12, London, 1876. Mounsey (Journey through the Caucasus, p. 329, London, 1872), who was in Baku July 18-19, 1871, writes of the temple as if of Zoroastrian origin, but speaks equally of the priest as a ‘Dervish from Delhi.’

[31] I have not been able to find any record of such missions in Patell's Parsi Prakash, Bombay, 1888. Nor do I know on what authority Orsolle (Le Caucase, p. 141, Paris, 1886) states that after the chief priest was assassinated in 1864 the Parsis of Bombay sent another, who was gradually forgotten and withdrew in 1880, although the latter date is about correct; cf. pp. 64, 66, below. Orsolle presumes throughout (pp. 140-141) that the temple was Zoroastrian, and refers to the profanation of the sacred fire in burning the bodies of the priests that died in service.

[32] O'Donovan (Merv Oasis, 1.37-39), who visited the fire-temple in 1879, speaks of it as ' of ante-Mussulman days ' and mentions 'Guebre worship.’ Furthermore, an article 'by a visitor,’entitled' An Ancient Zoroastrian Fire-Temple at Baku," in the magazine Men and Women of India, vol. 1, no. 12, p. 696 (Bombay, Dec. 1898), presumes throughout that the sanctuary was 'Parsi ' and says that ' thirty years ago… the last Zoroastrian attendant disappeared. 'A photograph of the shrine is given.

[33] The four lines of the Persian tablet (XIII), whose last line is metrically imperfect, are in praise of fire, and read: 'A fire has been drawn up like the array of a mountain, | Who can reach up to its crest ? | "May the New Year of the abode be blessed " - he said | The house has become radiant (? lit, light-spear) from it.’

[34] See Stewart, JRAS. 1897, p. 311, and cf. Cust, ibid. pp. 316-318. Fergusson, History of Indian and Eastern Architecture, p. 3164316, London, 1876, describes two temples at Kata Kangra, a larger and a smaller one, after Cunningham, Archaeological Reports, 6. 178 ff., but neither the description nor the woodcut (from a photograph, p. 316) are sufficiently comprehensive to give a fair idea whether there be a resemblance in style between these two and the Baku temple.

[35] Stewart subsequently (p. 316) changes his phrase 'much more modern' to ‘probably more modern,’ in speaking of the central structure. The inscription on its front now proves it to be younger than the surrounding walls with their inscribed cells. Eichwald (pp. 182-183) implies that it did not exist in Gmelin's time (Gmelin, Reise, 3. 45), but the latter work is not accessible to me for reference, Stew art's further statement (p. 312), that this middle building ' did not contain the fire," must be understood to refer to the especial occasion of his visit, because the pipes for the naphtha still remain in the flooring and on the roof. The account by Ussher, p. 207, more-over, shows that gas burned from the hole in the central edifice, and the frontispiece to his book shows the flame there as well as blazing from the roof.

[36] Brugsch, Im der Sonne, 2 ed. p. 66, Berlin, 1886. Orselle (Le Caucase, p. 140) uses the same phrase, ' dans le style indien," yet see above, p. 62, n. 2, end.

[37] Jackson, Persia, 254, 342. For a memorandum of the style of the Indian temple at Kangra, see above, p. 54, n. 1.

[38] So also Abbott, JAOS. 29. 303.

[39] Hanway, Caspian Sea, 1. 377 = 3 ed. 1. 260 Eichwald, ; Reise, 1 . 183, 217.

[40] See Cust in the article by Stewart, JRAS. 1879, p. 316, and compare Henry, Baku, 26. p.