Iconographic and Linguistic Interpretations of the Sasanian Clay Bullae in the Framji Dadabhoy Alpaiwalla Museum, Mumbai

Abstract

This article presents the first comprehensive study of six unprovenanced Sasanian clay bullae from Iran, held at the Framji Dadabhoy Alpaiwalla Museum in Mumbai. Bearing impressions from personal, official, and administrative seals, the seal impressions are here analysed for their iconography, stylistic characteristics, and Middle Persian inscriptions. The study proposes a chronological framework for the seals that produced their impressions on the bullae and situates the visual motifs within broader Sasanian artistic and symbolic traditions. It explains how these seal impressions reflect the intersection of visual culture, administrative function, and personal identity in Sasanian society, highlighting the complexity of interpreting sigillographic material where artistic convention and institutional roles converge. The inscriptions have been deciphered and enhance our knowledge of administrative practices in the Sasanian Empire. In particular, the study revisits the much debated term ⟨mgwh⟩ within the context of Sasanian administrative epigraphy.

1. Introduction

From 13th November to 5th December 2024, Moradi and Hintze conducted a curation and documentation project at the Framji Dadabhoy Alpaiwalla Museum in Mumbai. The project focused on archaeological materials from Iran preserved in the museum’s archives and was a preparatory effort for the forthcoming museum’s official reopening to the public (Harper 1973a)1. During this period, we documented a collection of 323 coins spanning a chronological range from the Achaemenid to the Afsharid periods. We are currently cataloguing the coins for future publication.

In addition to the coins, the museum’s collection includes one Sasanian intaglio, two seal stones and six clay bullae. The latter constitute the focus of this article. The inventory numbers for the six Sasanian bullae range from 214 to 219 in the museum’s catalogue. The exact provenance of these bullae remains uncertain, but the museum’s register indicates that these artefacts were brought from Iran to India by Dr. Jamshedji Maneckji Unvala (1888–1961), a Parsi scholar-priest specialising in the archaeology of Iran, ancient Indo-Iranian languages, and Zoroastrianism.

As a member of the French Archaeological Expedition at Susa from 1927 to 1939 (De Mecquenem 1980, pp. 35–47), Dr. Unvala amassed a significant collection of artefacts, which he later donated to the Alpaiwalla Museum. There is no record indicating when he took the objects to India, but this must have occurred during his participation in the excavations at Susa. These artefacts were reportedly obtained from archaeological sites such as Susa and the ziggurat of Chogha Zanbil in Khuzestan province, as well as from an unknown site in the district of Nehavand in Hamadan province. They include seals, bullae, potsherds, clay figurines, glazed decorative bricks, and inscribed bricks and stones, etc. Although no direct evidence links the bullae to Susa, the fact that the French expeditions at Susa uncovered several bullae within the Sasanian stratum on the Acropolis and the Royal City (Casanova 1925; Unvala 1944, 1953, p. 56), suggests that they may have come from there.

This article investigates the Alpaiwalla Museum’s collection of Sasanian bullae and seal impressions. This small corpus offers valuable insights into the administrative practices, iconographic conventions, and socio-cultural dynamics of the Sasanian Empire. Through a stylistic and iconographic analysis of the motifs on the seal impressions, we propose a chronological framework for the bullae and discuss the cultural and symbolic significance of their imagery. In addition, we analyse the Middle Persian inscriptions preserved on several seal impressions, offering new interpretations where legible and situating them within the broader context of Sasanian epigraphy.

2. Bulla inv. no. 214 (Figure 1)

Bulla inv. no. 214 measures 38 × 46 × 17 mm and weighs 32.07 g. Both the obverse and reverse are light brown in colour. The reverse is relatively flat, with slightly raised edges, and features a double crosswise string hole for passing a cord or string to bind the object it once sealed. A small part of this bulla is broken off, and what remains bears five seal impressions. Frye (1970, pp. 83, 436 fig. 14) reports that he saw this bulla in “the Parsi Museum (Kharegat Hall) in Bombay”. Publishing its administrative seal impression no. 1, he provides a reading of it, but neither comments on the bulla itself nor on the other four seal impressions it bears.

Figure 1. Bulla inv. no. 214.

Figure 1. Bulla inv. no. 214.

2.1. Bulla 214, Seal Impression 1

At 9 o’clock, an impression, 18 × 18 mm, of a circular seal with a flat engraved face appears. The seal impression displays no image and only text. The text consists of three lines at the centre and one line in the margin, all in Middle Persian lapidary script (Figure 2).

Frye (1970, p. 83) reads the inscription as follows:

At the centre:
Line 1 ⟨n]ywš[x]⟩	Nisha-
Line 2 ⟨p[w]xl⟩	pur
Line 3 ⟨mgwx⟩	magūkh(udā)
In the margin:
⟨ʾplštl⟩	Abaršahr

Later, Gyselen (2019a, pp. 23–26, 431, ATb256a) published a drawing of this seal impression and revised Frye’s transliteration, transcription, and translation as follows:

Centre:
Line 1 (starting at ca. 2 o’clock)	⟨[n]ywš[h]p⟩
Line 2 (starting at ca. 3 o’clock)	⟨[w]hl⟩
Line 3 (starting at ca. 4 o’clock)	⟨mgwh⟩
Margin (From ca. 6 o’clock to ca. 1 o’clock) ⟨ʾplštl⟩
Lines 1–3	/nēw-šābuhr mowūh/
Margin	/abaršahr/
‘Mowūh de Nēw-Šābuhr, (province de) Abaršahr’

Gyselen (2019a, pp. 23, 431) shows two annulets preceding and following the inscription in the margin through drawing (Figure 3). However, a close examination of the seal impression on the bulla itself suggests that the second circle in the drawing could represent the letter mēm ⟨m⟩, although its placement does not align with the expected content of the inscription. However, some other Sasanian seals and seal impressions do bear single letters as abbreviations for toponyms, some of which remain to date unidentified. The possibility cannot be excluded that the mēm ⟨m⟩ on this seal impression serves a similar function. Regarding the third circle at 9 o’clock in Figure 2 above (not represented in Gyselen’s drawing reproduced in Figure 3 below), it is tempting to interpret it as final ⟨y⟩ for the word ⟨nywšhphl⟩ and read it as ⟨nywšhphly⟩, a form attested on other seals (Gyselen 2019a, p. 423, ATb42a). However, the distance between ⟨l⟩ and the supposed ⟨y⟩ is too great to support this interpretation. The circle which in Gyselen’s drawing precedes the first letter (aleph) of the name ⟨ʾplštl⟩ written in the margin, should instead be represented as a rectangle, as this is the form it has on the seal impression (Figure 2).

Apart from the interpretation of the circles in the margin, there is no ambiguity concerning the reading of the inscription in the margin. Some uncertainty, however, surrounds the reading of the characters at the centre because some letters in line 1 are damaged. This makes it difficult to read them with confidence. Gyselen’s drawing (above, Figure 3) shows all characters preceding ⟨š⟩ as being lost, although she transliterates the first line at the centre as ⟨[n]ywš[h]p⟩ (Gyselen 2019a, p. 431, ATb256a). Her restoration implies that the engraver has omitted the character ⟨h⟩ before ⟨p⟩.

The new images we have been able to take of the seal impression confirm Frye’s and Gyselen’s restoration of the first letter in line 1 in the centre as nūn. The following sign can be recognised as yōd. The third letter appears to be a wāw which is placed above the lower horizontal line of the character yōd. While the last letter of line 1 is a clear ⟨p⟩, the preceding characters are not entirely clear due to damage to the clay after the characters ⟨nyw⟩ (Figure 4).

Although there is no question that lines 1–2 at the centre attest the place name Nēw-šābuhr, it is unclear whether three or four vertical lines precede the ⟨p⟩ at the end of line 1. In the former case, the first two vertical lines would belong to the character ⟨š⟩ and the third one would be the tail of a ⟨p⟩. Accordingly, the wāw would be followed by a gap before the šīn, similar to another seal bearing the name Nēw-Šābuhr (Gyselen 2019a, p. 23, ATb42a). It is not unusual for a compound word to feature a gap between its two components. However, the spelling of the name Šābuhr with ⟨šp-⟩ rather than ⟨šʾp-⟩ or ⟨šhp-⟩ is extremely rare, if attested at all. Gignoux (1986, p. 161, no. 858) lists a single and uncertain instance in the form ⟨š]pyhry⟩ in TS18 (Göbl 1976, p. 44, Tafel 63/85). Apart from the first character being restored, the spelling of the name with yōd rather than wāw also renders Gignoux’s identification of that name in TS18 as doubtful. In the absence of any clear evidence for the spelling of the name Šābuhr with neither hē nor aleph in the first syllable, one would have to assume that the seal cutter accidentally omitted one of these characters. Gyselen (2019a, p. 431, ATb256a) accordingly adds a hē to her transliteration of the name. However, there is little justification for the omission of a character considering the word space after ⟨nyw⟩. It therefore seems worthwhile to consider the option of seeing four vertical lines preceding the ⟨p⟩ at the end of line 1. The first vertical line is partly damaged, but the character ⟨š⟩ can be recognised, followed by ⟨ʾ⟩ and ⟨p⟩, which would have no tail. There is no uncertainty about the reading of lines 2 and 3. We therefore transliterate the inscription in the field as follows:

Centre:
Line 1 (starting at ca. 2 o’clock)	⟨nyw šʾp⟩
Line 2 (starting at ca. 3 o’clock)	⟨whly⟩
Line 3 (starting at ca. 4 o’clock)	⟨mgwh⟩
Margin (from ca. 6 o’clock to ca. 1 o’clock)	⟨ʾplštly⟩
Lines 1–3	/nēw-šābuhr mowūh/
Margin	/abaršahr/
‘Mowūh of Nēw-Šābuhr, (in the province of) Abaršahr’

Although the standard spelling of the name Šābuhr is with hē in the first syllable, spellings with aleph are also found (Gignoux 2003, p. 61, nos. 313 and 314). There is no need to assume that the engraver omitted ⟨h⟩ after the aleph in the first syllable of the name, as spellings with both aleph and hē are only rarely attested as, for example, in Burzād-Šābuhr ⟨bwlcʾt šʾhpwhly⟩ (Gignoux 1986, p. 61). Similarly in New Persian, the name appears both with and without ⟨h⟩: Šāhpur (شاهپور) and Šāpur (شاپور). The city of Nēw Šābuhr is mentioned in the Pahlavi text Šahrestānīhā-ī Ērānšahr §15 (Daryaee 2002, pp. 18, 39). Gyselen (2019a, p. 134) describes it as the capital of the province of Abaršahr.

Frye (1968, pp. 120–28; 1970, p. 83) transliterates the word in line 3 as mgwx and transcribes and reconstructs it as magūkh[udā]. Later he transliterates and interprets it as mgw-x(wtʾy) ‘chief priest’ (in a town or district), whom he considers to have been in charge of the town archives (Frye 1973, p. 50). He refers to the title ⟨ʾtwry ZY mgwxwdʾ⟩ ‘fire of the Magi-chief’, which Horn (1890, p. 673, no. 125, pl. IIb) had read on a 5th-century chalcedony seal held in the British Museum. Bivar (1969, p. 51, no. 119718, pl. 3, AE4), however, reads it as ⟨ʾtwlyxwdʾt ZY mgwxwdʾt⟩, while Gignoux (1987, p. 21) argues that this reading is misleading and rejects the existence of the word ⟨mgwxwdʾ⟩ on that seal.

The word ⟨mgwh⟩ has been widely discussed by scholars and continues to be subject to debate (Dang 2022, pp. 84–89). Maricq (1959, pp. 267–68) interprets the final ⟨h⟩ of the word ⟨mgwh⟩ as the initial character of the word ḥʾnk, the Persian word khāna, meaning ‘house’. He reconstructs the term as mgw-ḥ(ʾnk) (lit. ‘mgwh-house’) and translates it as the ‘bureau des mages’ or ‘office of the magi’.(Unvala, 1953)2. However, abbreviations of this type are unknown in Middle Persian texts. Frye (1960, p. 33) reports that Mordtmann and Horn read the term as magū ZY but that Herzfeld would have pointed out that the last letter of ⟨mgwh⟩ never take the usual form of ZY but almost always appear as ⟨h⟩. Frye refers the reader to Herzfeld apud Wilkinson (1936, p. 180), but in that article Herzfeld puts forward no such argument. Herzfeld assisted Wilkinson in reading the inscriptions, as Wilkinson (1936, p. 180, fn. 4) acknowledges, and transcribes the word as maguh. Without making any attempt to transliterate or transcribe the word and comment on its formation, Herzfeld (1938, p. 417) suggests that this term denotes ‘the office of the mōbed, the qāḍi’ (judge).

Gropp (1974, pp. 137–39) argues that mgwh refers to an institution and traces the word back to Old Iranian *magupaiθya, through intermediate stages such as *magūϑya > magūh (mgwh). He suggests that the mgwh institution was similar to the Parsee punchayet in Bombay. However, this idea is rather anachronistic as the latter is commonly thought to have started only in the 17th century and in the specific context of the Parsi community of India (Hinnells 1989). According to Gropp, mgwh designated a ‘Magi administration’ that evolved during the second half of the Sasanian dynasty, when a hierarchical territorial-administrative structure—village, district, province—was firmly established.

Gignoux (1974, pp. 171, 179, 180, 186; 1978, pp. 6, 16, 17; 1985, pp. 199, 205, 206, 210, 212, 213; Gignoux and Gyselen 1987) transliterates the term as mgwḥ and transcribes it as mogvēh/mog-vēh, interpreting mgwḥ as an institution—the ‘bureau des mages’. Gignoux (1984, p. 560, fn. 20; 1987, pp. 22–23) interprets the postulated compound mog-vēh as meaning ‘mage savant, instruit’, in English ‘learned mage, educated’, on the grounds of the alternative spelling ⟨mgwyd⟩, of which he considers ⟨mgwh⟩ to be a simplified spelling. He connects what he transcribes as vēh with the old Iranian verb vid ‘to know’ (Gignoux 1981, p. 34) and suggests that mgwḥ referred to an educated priest, probably below the mowbed but above the simple mgw, and responsible for the religious administration of a village or a small town, as evidenced by the seal impressions on bullae.

However, as Khurshudyan (1987, p. 94) points out, aside from orthographic difficulties, the main problem with Gignoux’s interpretation is that there is not a single seal—whether gem or bulla—where mgwḥ is accompanied by an anthroponym, while numerous seals bear the word mgw /mow/ along with a personal name. Khurshudyan (1987, pp. 93–94) argues that mgwḥ is an appellative derived from mgw, and, because it is represented in different parts of the Sasanian empire—as evidenced by numerous seals bearing this term—it must have referred to an institution that existed throughout the territory of Iran at the provincial, city, and district levels. Khurshudyan (1987, p. 95) reports about the view of Livshits, who interprets the spelling ⟨mgwḥ⟩ as mow(w)ah, meaning ‘abode of the magi’, and derives it from a postulated earlier form mow(w)ahah, with haplology of -ah. Khurshudyan reads the word as mow(w)ih and suggests that it may have referred to a gathering place for magi, similar to the Persian khānaqāh among Sufis. Since Zoroastrians did not have monastic institutions, he surmises, the term may have come to denote a collective body of local magi, and that this would be the meaning of the word on Sasanian seals.

Sundermann (1989, p. 362) also rejects Gignoux’s analysis of the word and instead offers the reading mowī. According to Sundermann, the word might derive from an Old Iranian *maguθwa- ‘office of the magi’, comparable to the Avestan aŋhuϑwa- ‘function of ahu’ and ratuϑwa- ‘function, office of ratu’. Gyselen (2015) transliterates the term as mgwh and transcribes it as mowūh (Gyselen 2019a, pp. 289–93). While Gyselen provides no philological explanation for this transcription, her transliteration and transcription have gained general acceptance among scholars.

However, a reading mowīh may be preferable. Since the spelling ⟨mgwh⟩ is based on the historical spelling of the word ⟨mgw⟩, which was pronounced /mow/ at the time this administrative term was probably created, it may be worthwhile considering that the word is formed with the suffix -īh. It may have been pronounced /mowīh/ ‘office of the mows’. Such an interpretation could be supported by occasional alternative spellings ⟨mgwydy⟩ and ⟨mgwyd⟩ /mowē/, surveyed by Gyselen (2012a, pp. 128–30).

The term ⟨mgwh⟩ appears widely, but only in administrative seals and seal impressions from the late Sasanian and early Islamic periods. A small number of bullae, bearing the term ⟨mgwh⟩, from the Ṭabarestān archive were still attached to mid-eighth-century documents (Gyselen 2019a, pp. 213–15; 2019b, pp. 151, 154). Within the corpus of the Sasanian administrative seals, the impression of the seal of ⟨mgwh⟩ constitutes the highest number. According to Dang (2022, p. 90), this office appears on 21 surviving seals and in 423 seal impressions on extant bullae, suggesting a role tied to local administration at the level of a ‘district’, or ‘canton’ (tasōg) within a larger province. Recent discoveries of additional bullae have significantly increased these numbers.

The term mowūh is entirely absent from Zoroastrian literary sources (including the Sasanian Low Book), rock inscriptions and even Manichaean Middle Persian texts. Therefore, the meaning of this word is established solely from the inscriptions on seal stones and bullae. Despite ongoing ambiguities regarding its accurate transcription, meaning, and administrative role, numerous speculative interpretations have been proposed without much, if any, corroborating evidence. For instance, it has been defined as ‘bureau des mages’ (Maricq 1959, pp. 267–68; Gignoux 1987, p. 20), ‘[mgwḥ] désignant la communauté locale, tout comme dans une paroisse, catholique’ (Gignoux 1991, p. 67), ‘the magi administration, an institution probably not differing much from the Parsee punchayet in Bombay’ (Gropp 1974, p. 138), ‘council of the magi’ (Gubaev et al. 1996, pp. 55–57), ‘the office of the Magi’ (Wiesehöfer 2001, p. 188), and ‘office of the moγ’ (Cereti and Bassiri 2016, p. 15). Gyselen (2019a, pp. 296–97) believes that the translation ‘bureau of the magi’ proposed by most authors is appropriate for the collective nature of the administration at the district level of the mowūh. She further hypothesises that the local or regional administration of the mowūh operated under the authority of a higher official, such as the mowbed or drīyōšān jādaggōw ud dādwar, although she acknowledges the lack of direct evidence for such a relationship. Similarly, Wiesehöfer (2001, p. 188) argues that the mowūh, exclusively functioning at the district level, handled the settlement of disputes locally and thus served as a mediating authority for the offices of the mowbed or drīyōšān jādaggōw ud dādwar, both of which operated at the provincial level. Dang (2022, p. 95) suggests that the mowūh administration may have functioned independently of the mowbed at the local district level, with the mowbed possibly intervening only in specific cases or legal matters.

2.2. Bulla 214, Seal Impression 2

At 3 o’clock, an impression, 18 × 18 mm, of a circular seal with a flat engraved face appears. It features a standing male figure shown in profile, facing to the right (Figure 5). The head is oval-shaped and proportionate to the body. No clear indication of head covering is discernible due to serious damage. His hair is depicted hanging from the back of the head, although the details are somewhat eroded and unclear. Despite the severe deterioration, the face retains traces of a prominent short nose, a large fully opened eye with prominent eyelid, and full lips. The beard is very faint, giving only a general impression of the length rather than detailed characteristics. The eyebrows and ears are not visible.

The right arm of the figure is bent at the elbow and extended forward, with the thumb and forefinger bent towards one another while the other three fingers point upwards. This auspicious gesture could be interpreted as an expression of greeting, admiration, reverence or invocation. The left arm is slightly bent and horizontally extended forwards, holding a circular string of ten large, round beads, likely to represent pearls. The feet are proportionately small and thin in relation to the body. They are shown in profile and positioned awkwardly in opposite directions. It is uncertain whether the figure wears footwear or is barefooted. Although a vertical line extending downwards from the left heel could indicate some sort of shoe, this feature cannot be taken as conclusive evidence, as it may simply represent an overemphasised rendering of the foot’s heel.

The figure is dressed in a long, belted tunic that extends to the ankles, featuring a smooth and simple texture that conveys a sense of formality. The tunic is form-fitting with sharply defined edges, flaring slightly at the bottom to suggest movement or a stylised design. It has a round collar with a prominent neckline but lacks elaborate detailing. Two long strings, serving as a looped belt, extend from the figure’s waist and hang vertically down the front, ending in tassels shaped like bird heads. To the figure’s left, there are a crescent moon and a six-pointed star, which might indicate celestial or religious significance.

The pose of a standing figure raising their right hand, with the thumb and forefinger bent toward one another, either holding a flower or empty-handed, is a common motif in Sasanian seals and, less frequently, on rock reliefs. For instance, on several seals, the goddess Dēn is depicted extending her right hand in this gesture while holding a flower, likely a lily, which is a specific attribute associated with her (Gnoli 1993). A similar gesture appears on the rock relief at Barm-i Dilak, located 10 km southeast of Shiraz in Fars province, where the Sasanian king Bahrām II presents a flower to the queen (Shahbazi 1998; Overlaet 2010, pp. 339–40). Several seals also portray unidentified standing male and female figures performing this gesture while holding a flower in their hand (Bivar 1969, pp. 63–64, pl. 7, CC 1–13; Brunner 1979, p. 40, fig. 8; Gignoux and Gyselen 1982, p. 36, pl. III, 10.1; Gyselen 1993, p. 77, pl. I, 10.A.4; 2007, pp. 96–97, 10.2–10.3). By contrast, other seals depict male figures seated on a chair or bench (Gyselen 1993, pl. III, 10.B.18, 10.B.19) or a female bust shown in right profile (Gignoux and Gyselen 1989, pl. II, 16), all holding nothing with the thumb and forefinger bent toward each other.

The above examples demonstrate that the gesture of extending the right hand, with the thumb and forefinger bent toward one another, either holding a flower or empty-handed, was performed by a deity, a king, or a person, male or female. Although the exact identity of the figure on the seal impression from the Alpaiwalla Museum is unknown, given the content of the accompanying inscription, it is likely that the figure represents a priest in a ceremonial context. A well-known parallel can be found in the rock reliefs of Kerdīr, the high priest of the early Sasanian period, who is shown raising his hand in a gesture of reverence, though with notable iconographic differences. In the relief at Naqš-e Rajab, Kerdīr’s forefinger is bent and points upward, while at Naqš-e Rustam his thumb is directed upward. By contrast, on the Alpaiwalla bulla the hand is extended forward, with the thumb and forefinger bent toward each other, indicating a distinct variant of the reverential gesture.

This seal impression is of the same size as that of the administrative seal, alongside which it appears on bulla 214. Its size and position on the bulla as well as being both iconic and epigraphic suggest that it is the official seal of the cosignatory, while the three smaller seals impressed on the edge of the bulla are those of the witnesses who were present at the sealing process. In the margin of seal impression 214.2, one line of an inscription in Middle Persian cursive script surrounds the standing figure. The letters are partly worn, but the word ⟨mgw⟩ can be recognised at 5 o’clock on the circumference (Figure 5). Although part of the third letter has disappeared, one can still discern the traces of a complete waw, and the gap following the character suggests that the word ends here. A similar shape of the character wāw is found in the word ⟨mgwpt⟩ (Figure 6) on a seal impression from Qasr-i Abu Nasr (Frye 1973, p. 60, sealing D.1, pl. I.1 and drawing). The drawing of that seal impression had previously published by Wilkinson (1936, p. 181). The presence of the word ⟨mgw⟩ on seal impression 214.2 further confirms the identification of this seal impression as that of the cosignatory. The impression of the administrative seal next to it on this bulla indicates that the cosignatory belongs to the mowūh administration of Nēw-Šābuhr of the provincial administration of Abaršahr.

Gyselen (1995, pp. 124, 143–48), who surveys the inscriptions found on seals and seal impressions of magi, concludes that about 80% of them follow the structure of the personal name, the function (mow) and the father’s name, while ca. 16% omit the patronymic. A small number of inscriptions provide a personal name, the priestly function mow, and a pious formula (2%) or a personal name, mow, patronymic, and a pious formula (2%). Accordingly, one of these four patterns is expected to be present in the inscription found on seal impression 214.2 of the Alpaiwalla Museum. Considering that the characters ⟨mgw⟩ appear at 5 o’clock, the inscription should start at 11 o’clock. The area immediately to the left of the head of the standing figure is damaged and the surface of the seal impression is lost. There seem to be traces of two characters, the first of which could possibly be aleph, sāmekh or šīn and the second perhaps a cursive ḥēth/hē (Figure 7), although, due to the damage to the surface, it is not entirely certain that the traces to the left of the head are letters at all. Assuming that they are letters, possible readings could be /sah-/ or /šāh-/. Gignoux (1986, p. 156, no. 822) records a personal name Sahād, whose first syllable he explains as an abbreviated name formed from the present stem sah- ‘to seem proper’. Alternatively, the first part of the name could be the word šāh ‘king’. This word is present in the hypochoristic name Šāhag, which, however, is spelt with aleph and hē in the first syllable (Gignoux 1986, p. 163, no. 866). The following character looks like pē followed by a cursive lamed and a nūn, thus possibly providing the reading ⟨pln⟩ /farr/, although this also remains uncertain. The next letter looks like another cursive lāmed. The following character could possibly be read as a tāw. The traces which follow show two short vertical strokes and a fat horizontal line at the bottom. Possibly the first vertical stroke and the horizontal line represent a gimel followed by a cursive wāw or, possibly, a cursive nūn or rēš. The next letter has disappeared, but traces are present in the form of a line which is bending to the left at the top and bottom. This is followed by a second curved line of about the same shape. The shape of the two curved lines could point to the characters ⟨s⟩ or cursive ⟨š⟩, although the second curved line could also be interpreted as representing ⟨b⟩. However, that would leave the preceding first curved line unexplained.

Based on these considerations, the resulting reading of the characters to the left of the standing figure is ⟨šhplnltgws⟩. These characters should represent the personal name of the owner of the seal, possibly starting with šāh ‘king’ followed by farr- (< Old Persian farnah-), which also occurs, for example, in farrbay, a compound of farnah- ‘glory, light’ and baga- ‘god’. The next part of the name could be /rad/ (Avestan ratu- ‘master, judge’), which also appears, for example, in Mānrad (Gignoux 1986, p. 116, no. 564) and Bayānrad (Gignoux 1986, p. 56, no. 196). The final element ⟨gws⟩ or, perhaps more likely, ⟨gwš⟩ according to the reading presented here, would turn the name into a four-part name like, for example, Ādur-Farrbay-Gušnasp (for this and more examples of three-, four- and five-part names, see Gignoux 1986, p. 35, nos. 55–57, p. 36, no. 67, p. 56, no. 194, p. 81, nos. 344–47, p. 88, no. 384, p. 97, no. 440, p. 108, no. 514, pp. 108–9, nos. 514–16, p. 111, no. 534, p. 183, no. 1007, etc.). While no interpretation for the reading ⟨gws⟩ /gōh?/ emerges, the reading ⟨gwš⟩ could represent MP gōš ‘ear’ (OP gauša-), which Gignoux (1986, p. 55, no. 189) considers to be present in the name Bargōš and for whose first part he considers a Semitic origin. Another word for ‘ear’, ōš (Av. uš- ‘ear’) appears in the name Ōš (Gignoux 1986, p. 140, nos. 716–19) and in Māh-ōš, which Gignoux (1986, p. 113, no. 544) translates as ‘qui a l’intelligence de Māh’ and in Warm-ōš (Gignoux 1986, p. 175, no. 954). Accordingly, the name on the bulla under discussion would read šāh-farn-rad-gōš ‘who has the glory of a king (and) who has the listening of a judge’.

To the right of the feet of the standing figure the characters following the word ⟨mgw⟩ /mow/ are expected to represent the name of the father of the mow (Figure 8). The first character looks like bēth. Then follows a character which could be another bēth written above the horizontal line of the preceding character, or it could be the first part of an aleph. If bēth is read, then the short vertical line visible after it could be the first part of an aleph and the next vertical line would be its second part, with its horizontal line written further down. However, considering that the horizontal line of the second part of an aleph is expected to be written at the same hight as the first part, it seems preferable to interpret the second character as an aleph rather than as a bēth. The lower part of the third character looks like zayin. However, since the upper part of the character is a straight vertical line rather than a line bending to the right, it is more likely to be a bēth. This results in the reading ⟨bʾb⟩ /bāb/, a name also found elsewhere in Sasanian inscriptions (Gignoux 1986, p. 52, no. 269; 2003, p. 25, no. 52). The next character, letter 4, could be a gimel, followed by a vertical line, which could possibly belong to a cursive wāw, rēš or nūn. The next letter 5 can be read as a cursive tāw and this is followed by what at first sight looks like an inverted dāleth. This, however, is unlikely. The letter seems to have a horizontal line in the middle which joins the curved line coming up from the bottom. This could suggest a cursive mēm. However, for a cursive mēm one would expect some traces of a horizontal line joining onto the curved part on the left of the character. It therefore seems more likely that what appears to be a horizontal line at the centre of the character in fact belongs to the upper part of the character. Upon closer inspection of the bulla itself and with the help of multiple photographs taken from different angles and with different lighting, it appears that the line of the character forms a three-quarter circle open at the upper right, where traces of the line going up are still present (Figure 9). This suggests the character pē.

The next character, letter 8, has similarity with a cursive pē, but since it has traces of a tail at the upper left, a tāw seems more likely, the small tail being what is left of the vertical line to the left of the round body of the tāw. The last letters of this part of the inscription are damaged. It seems that what is present are the remains of three letters, letters 9–11 (Figure 10).

The first of them, letter 9, shows a curved line open to the left and a second line to its left of the same shape. Although the character is damaged at its left part, this could represent a sāmekh ⟨s⟩. What follows seem to be faint traces of an aleph. Its right part is damaged, but its left part is visible. The horizontal line at the bottom of the aleph joins up with the last character, a vertical line that probably represents a cursive nūn. These considerations produce the following reading of the inscription:

Starting at 11 o’clock ⟨šhplnltgws mgw bʾbgwtptsʾn⟩

/šāhfarn-rad-gōš mow bāb-gōtaftasān/

‘Šāh-farn-rad-gōš(?), the priest, son of Bāb-Gōtaftas(?)’

While ⟨bʾb⟩ /bāb/ is well attested in Sasanian personal name, the remaining part of the name as read above remains obscure. The element -gō- could belong with OIr. *gau̯- ‘cow, bull’. It is attested in the Middle Persian names Gōbārān (Gignoux 1986, p. 89, no. 390) and Gōbed (Gignoux 2003, p. 38, no. 145). The element -tafta- could be the verbal adjective of tap ‘to be, become hot’, but it is hard to make sense of it in combination with gaw- ‘cow, bull’. The ending -ān of the final element -sān or possibly -hān could be the suffix marking a patronymic, but the analysis and meaning of a name Gōtaftas or Gōtaftah remain obscure.

2.3. Bulla 214, Seal Impression 3

At 1 o’clock on the edge of bulla 214, an oval-shaped convex seal has been impressed, measuring 8 × 10 mm. The seal impression, which has no inscription, shows a lion walking towards the left, with its head shown in three-quarter profile (Figure 11). The lion’s snout is beautifully executed, its mouth is open, and its eye is glaring. The mane is indicated by a series of sparse oblique lines, covering the head and neck. The lion has a heavily muscled body, and its legs are relatively short. The paws and claws are large in proportion to the body. The backside of the lion is depicted as the head of an unidentified animal facing right, while the snout, eye, and head are well executed. Its ears are located on the sides of the head and slightly pointed in shape. Coming out of the mouth of this animal, the tail of the lion is shown curving upwards, continuing above the body and ending in a tuft arranged in the shape of a bird. From ca. 2 to 11 o’clock, there are traces of an unidentified object.

2.4. Bulla 214, Seal Impression 4

At 2 o’clock, an impression, 10 × 11 mm, of an oval-shaped convex seal appears (Figure 12). The seal impression is identical to seal impression no. 3 (above, Section 2.3) on the same bulla. Occasionally, though not frequently, one or more of the small seal impressions are repeated on Sasanian bullae. The purpose of the repetition remains unclear. Frye (1968, p. 119) considers it to be unlikely that the repetition has any significant meaning. In his view, the repetition reflects the desire of particular witness to fill empty space or to ensure the impression of their seal was clearly recognised, given the small size of the impressions. However, it is difficult to accept Frye’s hypothesis, considering that even with the repetition, ample empty space remains. If the repetition’s purpose was to fill up blank space, why were the seal impressions not repeated more than twice to completely fill the area? Furthermore, if the repetition was intended to ensure clear recognition of the seal, why were other small seals not repeated in the same manner? Although the presence of small seals may plausibly indicate the involvement of witnesses or a partnership in a commercial transaction, there is no compelling explanation for their repeated application. It is possible that such repetition served to reinforce the validity of the transaction or to emphasise the authenticity of its contents.

2.5. Bulla 214, Seal Impression 5

At 6 o’clock, an impression, 4 × 4 mm, of a circular convex seal appears. The actual seal stone was set in a lozenge-shaped mount of a pendant (6 × 10 mm) with its bale 3 mm long, giving the entire impression a drop-like shape. The seal impression exhibits a bird standing, facing to the right (Figure 13). The bird has a compact, rounded body, with a short neck and broad, relatively flat chest. Its head is round and smooth, with a flat, wide bill that is slightly pointed at the tip. The bird has short, sturdy legs positioned towards the rear of its body. Its body is covered with feathers, indicated by a series of sparse oblique lines.

3. Bulla inv. no. 219 (Figure 14)

Bulla inv.no. 219 measures 39 × 40 × 20 mm and weighs 31.53 g. The colour on the obverse includes dark grey and brown, while the reverse is dark grey and black. On the back, there is a crosswise slit through which a strap of leather or parchment was passed to secure the bulla to the object sealed. This bulla bears four seal impressions.

Figure 14. Bulla inv. no. 219.

Figure 14. Bulla inv. no. 219.

3.1. Bulla 219, Seal Impression 1

At 6 o’clock, an impression, 18 × 18 mm, of a circular seal with a flat engraved face appears. The sealing is aniconic and bears two lines of text in the centre and one line of text encircling the circumference, all in Middle Persian lapidary script (Figure 15).

A similar example of this seal impression was published in the form of a drawing by Herzfeld (1938, p. 422, fig. 23). Herzfeld identifies two toponyms: (h)askēn-rōd (Askēnrōd) at the centre and Ahmadān (Hamadān) in the margin. Transcribing the text at the centre of the sealing as (h)askē/īn rōd, Herzfeld comments that this is “a place called after a river, situated in Ahmadān city; to be studied”.

Gyselen (2012a, p. 127, I/177; 2012b, pp. 138–39, ZR11) provides photographs of two bullae held in the Ahmad Saeedi collection and the British Museum bearing the same seal impression and transliterates and translates the entire inscription. Subsequently, Gyselen (2019a, p. 101, ATb285a) presents a drawing of one of these seal impressions. In the reproductions published by Herzfeld and Gyselen the word /kust/ ‘side’ can be read clearly, and in the sample held at the Alpaiwalla Museum the letters ⟨k⟩, ⟨s⟩, and ⟨t⟩ are also recognisable. Accordingly, the inscription can be transliterated, transcribed and translated as follows:

Centre:
Line 1 (starting at ca. 3 o’clock)	⟨ʾskynlwty⟩
Line 2 (starting at ca. 4 o’clock)	⟨mgwh⟩
Margin (starting at ca. 8 o’clock)	⟨ʾḥmtʾn k[w]st[y] ZY štldstn⟩
Centre Lines 1–2	/askēn-rōd mowūh/
Margin	/aḥmadān kust ī šahrestān/
‘Mowūh of Askēnrōd (in the province of) Axmadān, side of the provincial capital’

The spelling of the name of the city of Hamadān varies in the sources, ranging from Old Persian h-g-m-t-a-n-i-y /hagmatānai̯/ (loc.sg.) ‘in Ekbatana’ (DB §32, col.2, l.76, Schmitt (2009, p. 61; 2014, pp. 185–86)), Elamite Ag-ma-da-na, Akkadian A-ga-ma-ta-nu, Greek Ἀγβάτανα, Ἐγβάτανα, to New Persian Hamadān (Brown 1997, p. 80). The name is rendered as ⟨ʾḥmtʾ⟩ in Aramaic, and in Armenian sources as Ahmatan (Hübschmann 1897 p. 17). The spelling ⟨ʾḥmtʾn⟩ on the seal impression 1 of bulla 219 shows the same three characters at the beginning as the abbreviated spelling ⟨ʾḥm⟩ on coins with Middle Persian inscriptions indicating the mint place (Mordtmann 1865, p. 393; 1879, p. 117). Klingenschmitt (1980, pp. 20–21, fn. 3) confirms the old etymology (Kent 1953, p. 212) of the name as going back to Old Iranian *ha-gmata- ‘come together’.

Herzfeld (1938, p. 422, fig. 22) and later Gyselen (2012a, p. 127, I/176; 2012b, p. 136, ZR7; 2019a, pp. 101–102, ATb286a) published two seal impressions of the second seal of the mowūh of Askēnrōd in the province of Hamadān kust ī šahrestān. One is found on a bulla bearing five seal impressions and belonging to the Ahmad Saeedi collection (Gyselen 2012a, p. 127, I/177). The other example is one of four seal impressions on a bulla of the British Museum (Gyselen 2012b, p. 136, ZR7). Notable differences distinguish this second seal from the first. In particular, the inscription of the second seal reads counterclockwise from 1 o’clock and the toponym Askēnrōd is written in two lines, whereas the other one reads from 8 o’clock and the toponym Askēnrōd is written in a single line and has a final yōd in the form of a vertical stroke. These differences clearly indicate that the second seal impression was produced using a different seal stone. This suggests that the office of the mowūh of Askēnrōd had two distinct administrative seals. However, it remains unclear whether both seals were used simultaneously or in two different timeframes. If the latter is the case, it is uncertain which of the two seals predates the other.

3.2. Bulla 219, Seal Impression 2

At 10 o’clock, an impression, 12 × 15 mm, of a circular seal with a flat engraved face appears. This seal impression exhibits two overlapping male busts accompanied by a line of inscription in Middle Persian cursive script on the circumference (Figure 16). This seal impression is that of the co-signatory of the transaction.

The bust has a static heraldic form and does not create any sense of motion through duplication. Only about half of the right bust is visible, with the rest covered by the left bust. Since both busts are identical, we limit our description to the left bust, which is fully depicted. The bust is shown in right profile with his shoulders and chest frontally in the impression. His facial features are detailed, with a long straight nose, prominent eyebrow, a large fully opened eye with defined eyelids, a visible ear, and full lips. The moustache is indistinct, while his short beard is indicated by a series of sparse oblique lines. The hair emerges from the base of the headdress as three parallel locks, tied with two parallel hair bands, after which the locks widen into curls that hang down to the shoulder. Each lock ends in a circular shape. The bust’s drop-earring consists of two round pearls or beads, one small and one large, suspended one above the other from his earlobe. Due to significant damage, it is unclear whether the bust wears a necklace. He wears a domed shape kōlāh/kōlāf (headdress) of the nobility adorned with a row of palmettes, each with five leaves. The kōlāh is hemmed with a diadem tied at the back of the head, with two short ribbons trailing behind. He wears a round-collared cloak with a prominent neckline. The drapery is conventionally shown as six rows of parallel folds, running obliquely from the shoulders. These folds would continue across the chest, but their details are lost due to the deterioration of the seal impression. No further details about the cloak can be discerned due to severe deterioration. The bust stands on a curved base that terminates in foliage at either end. Severe damage makes it is uncertain whether additional decorative elements, such as flowers or stars, are present on the base, as is common on many Sasanian seals.

Although not very common, overlapping busts can be found in Sasanian sigillography. Interestingly, most, if not all known seals with this design are associated with Zoroastrian priests (mow). Two such examples appear on bullae from Qasr-i Abu Nasr, one of which belongs to a mow named Narseh and the other one with a fragmentary illegible inscription (Frye 1973, p. 61, D112, p. 64, D421). Another example is a seal impression of a certain priest on a bulla in the Mohsen Foroughi collection. Frye (1968, p. 125, pl. XXVII.20) reads the inscription on that seal impression as ⟨bʾp-mgw ʾs⟩. A further example is the seal of the mow Ādur-farr, son of Astōyēn, housed in the Bibliothèque nationale de France (Gignoux 1978, p. 27, pl. V. 3.8; Gyselen 1993, p. 106, pl. XVIII, 20. L.2).

While it is tempting to interpret the bust on the bulla from the Alpaiwalla Museum as representing the seal’s owner, the iconography lacks distinctive characteristics that would allow for such an identification. In a discussion of the relationship between the content of an inscription with the imagery on a seal, Gyselen (2017, pp. 87–88) proposes that the decoration around the base of the kōlāf on the bust may provide insights into the position or function of the seal’s owner. If this hypothesis proves accurate, the similarities between the row of palmettes on the crown and the hairstyle—particularly the parallel locks—and those on the impression of the seal of the ‘Dādohrmazd son of Narses the counselor of the priests’ from Qasr-i Abu Nasr (Frye 1973, p. 61, D103), as well as the 6th-7th centuries seals of the dar-andarzbed or ‘counsellor of the court’ (Gyselen 2008, pp. 14–15, 18–19, 28–29, 51–56), may suggest that the bust on seal impression 219.2 from the Alpaiwalla Museum belongs to the class of anderzbeds. However, unlike the dar-andarzbed seals, no reference to dar-andarzbed appears in the inscription of the seal impression at the Alpaiwalla Museum. Given that the inscription explicitly refers to a mgw, it is probable that the bust represents a priest. Nevertheless, the question of why the same bust design would be duplicated on a single seal remains unresolved. It is also tempting to assume that one of the busts might represent the son and the other the father. This hypothesis gains plausibility in light of the fact that the inscriptions on the seals with duplicated busts refer to both the son and his father.

The hairstyle, particularly the parallel locks, is amply attested on Sasanian seals, bullae, coins, silverwork, and rock reliefs dating as late as the 3rd century. However, the similarity of both the hairstyle and the floral motif around the base of the kōlāf of the bust in the seal impression from the Alpaiwalla Museum to those found on andarzbed seals suggests that the original seal stone that produced this impression likely dates to the 6th or 7th century. This proposed dating is further supported by the juxtaposition of this seal impression with that of the administrative seal of the office of the mowūh of Askēnrōd, which dates possibly from the 6th century or later. According to the 7th-century Sasanian Law-Book Mādayān ī Hazār Dādestān (MHD 93.5–9; Macuch 1993, pp. 593, 596, 600–1), the office of mowbed and its seal were first introduced into the Sasanian bureaucracy during the reign of Kavād (r. 488–496, 498–531 CE).

In the margin of the seal impression, from ca. 5 to ca. 8 o’clock, there is a line of inscription in Middle Persian cursive script, which reads as follows:

⟨ʾzʾtmlt’ Y mgw Y pylwcʾn hwmsʾy⟩

/āzādmard ī mow ī pērōzān humehāy/

Āzādmard the priest, son of Pērōz, the one of good greatness’

The name Āzādmard is also attested elsewhere (Gignoux 1986, II, p. 52). It is usually spelt ⟨ʾcʾtmlt’⟩, but occasionally ⟨ʾzʾtmlt’⟩, as can be seen on a chalcedony seal in a private collection (Gignoux and Gyselen 1982, pp. 130–31, no. 2.34, pl. XXI 40.22). What is transcribed above as pērōzān is also attested as a personal name (Gignoux 1986, p. 147, no. 760), but the context of the present inscription suggests that here the form is the patronymic of the well attested name Pērōz (Gignoux 1986, p. 147, no. 759). The engraver squeezed the last word, here read as ⟨hwmsʾy⟩, into the remaining space on the circumference of the seal stone (Figure 17). The reading is therefore not entirely certain. Keeping in mind that the script is a cursive form of inscriptional Pahlavi, the characters following the word ⟨pylwcʾn⟩ can be read as ⟨hwmsʾy⟩ /humehāy/ or ⟨hwmʾhy⟩ /humāh/. Gignoux (1986, p. 85) records a personal name Farrox-mehā̆y, formed with a hypocoristic suffix -ā̆y (Gignoux 1979, pp. 64–65) from meh ‘great(er), old(er)’. Alternatively, one could posit a compound of hu ‘good’ with māh ‘moon, month’, but such a compound is absent from Gignoux’s lists of personal names (Gignoux 1986, p. 211; 2003, p. 75).

3.3. Bulla 219, Seal Impression 3

At 8 o’clock, an impression, 5 × 11 mm, of an elongated, octagon-shaped convex seal, is found on the side of the bulla. The seal impression shows a lion standing, facing right, and no inscription (Figure 18). The lion’s mouth is open, and its tongue is visible. His snout is beautifully executed. While other facial details of the lion are indiscernible, the mane is represented by obliquely raised liners, covering the head and neck. The lion has heavily muscled shoulders and chest, but a relatively slender body. Its legs are relatively short yet extremely powerful, giving it the strength to sprint and pounce. The details of the paws are unclear due to deterioration. The tail is shown curving upwards, ending in a tuft of fur. At 12 o’clock, there is a crescent moon alongside an unidentified object.

3.4. Bulla 219, Seal Impression 4

At 12 o’clock, an impression, 5 × 8 mm, of an elongated, hexagon-shaped seal with a convex face appears. The impression portrays a figure standing facing left (Figure 19). The figure has a long beard and moustache, though other facial features are unclear due to damage. A ribbon is tied around the figure’s head, trailing behind. The figure’s hands are not visible. He is dressed in a long garment decorated with horizontal bands and wears high-heeled shoes. His trousers are indistinguishable. A short, raised vertical line appears in front of the figure.

4. Bulla inv. no. 215 (Figure 20)

This bulla measures 24 × 27 × 9 mm and weighs 4.97 g. The colour on both obverse and reverse is reddish brown. The reverse side of the bulla lacks a string hole, typically used for the passage of a cord or string to bind the sealed object. The absence of slots, holes or cord marks raises the question of how the bulla was affixed to the object it meant to seal. One can, however, argue that it was probably used to authenticate the sender’s identity rather than a document or an object.

It is interesting that at a later date, a seal was pressed onto a piece of grey clay that had been added to the reverse side of the bulla. The seal impression, measuring 16 × 16 mm, displays a male portrait bust in right profile. Although the image is severely damaged, certain facial features including the nose, eye, and forehead adorned with a diadem can still be discerned. It is evident that this seal impression was applied after the original creation of the bulla, and is most likely a modern reproduction, as it partially covers the now-illegible ink inscription.

Figure 20. Bulla inv. no. 215.

Figure 20. Bulla inv. no. 215.

On the front side, the bulla bears four identical impressions, each measuring 8 × 16 mm, made by an elongated convex seal (Figure 21). Each seal impression features two lines of inscription in Middle Persian lapidary script, accompanied by decorative elements: five raised dots above and a five-pointed star flanked by a plant on either side. The repetition of the same seal impression makes this bulla exceptional, as no other example of duplicated epigraphic seal impressions on a single bulla has been documented to date. In other cases, duplicated impressions on bullae are usually small, anepigraphic, and attributed to witnesses, whereas on this bulla, the repeated impression is the principal—and only—seal impression.

The inscription on this bulla reads as follows:

Line 1 ⟨yzd pnʾh⟩

Line 2 ⟨ylk pnʾh⟩

Line 1 /yazd panāh/

Line 2 /… panāh/

‘Protection of the sacred being, Protection of …’

At the end of each of the two lines the word /panāh/ ‘protection’ can be recognised. In line 1, it seems to be preceded by four characters ⟨n⟩, ⟨y⟩ or ⟨b⟩, ⟨g⟩ or ⟨z⟩, and ⟨d⟩. However, it is unclear whether the first character, which looks like ⟨n⟩, forms part of the inscription or whether it results from an impression of the rim of the seal stone. If the latter is the case, then the inscription on line 1 can be read as ⟨yzd pnʾh⟩ /yazd panāh/ ‘protection of the sacred being’. If what looks like ⟨n⟩ does form part of the word, a reading ⟨nyzd⟩ emerges, but this makes little, if any, sense.

In Line 2, the first character is also ambiguous as it is unclear whether or not the curved line at 4 o’clock forms part of the inscription. If it does, the first character could be ⟨s⟩. The second character looks like ⟨l⟩ and the third like ⟨k⟩ or, possibly, ⟨d⟩. But what could ⟨slk⟩ or ⟨sld⟩ represent? That the curved line at the beginning continues upwards toward the first sign on line 1 suggests that it was caused by the impression of the rim of the seal stone. If so, the first character could be read as ⟨y⟩ or ⟨g⟩ or, possibly, a small ⟨b⟩. This would produce the reading ⟨ylk⟩ or ⟨dlk⟩ or ⟨blk⟩, but no interpretation emerges from such a reading.

5. Bulla inv. no. 216 (Figure 22)

This bulla measures 35 × 38 × 19 mm and weighs 27.28 g. The colour of the obverse is light brown while the reverse includes light brown (7.5YR-6/4) and greyish brown (2.5Y-5/2). The back side shows two crosswise string holes, one in body and one on back, through which a string or cord passed to secure the object sealed. Traces of a sack or fabric, with a 5 mm hem, to which the bulla was affixed are also visible on the reverse.

Figure 22. Bulla inv. no. 216.

Figure 22. Bulla inv. no. 216.

This bulla bears three seal impressions:

5.1. Bulla 216, Seal Impression 1

At the centre, an impression, 8 × 16 mm, of an oval-shaped seal with a flat, engraved face set in a mount appears. The image on the seal impression has deteriorated to such an extent that the species of the depicted animal cannot be identified (Figure 23).

5.2. Bulla 216, Seal Impression 2

At 3 o’clock, an impression, 10 × 13 mm, of an oval-shaped seal with a flat, engraved face appears. It exhibits a winged horse (Pegasus) standing facing right, with its head lowered (Figure 24). This seal impression features two globular indentations at 3 and 9 o’clock, suggesting that the bezel had two prominent knobs on its edge. The knobs were not only decorative but also practical, as they prevented slippage while pressing the seal stone into the moist clay, ensuring even pressure for a clear impression. The knobs were also oriented the seal correctly during application (Lerner and Skjærvø 1997, p. 71; Moradi and Hintze 2020, p. 115) and acted as a lever to remove the seal from the clay bulla after the impression (Gyselen 2007, pp. 25–26).

5.3. Bulla 216, Seal Impression 3

At 9 o’clock, an impression, 7 × 9 mm, of an oval-shaped seal with a flat, engraved face appears. It exhibits a predator animal walking towards the left (Figure 25).

6. Bulla inv. no. 217 (Figure 26)

This bulla measures 23 × 24 × 14 mm and weighs 5.75 g. Both the obverse and reverse are reddish brown in colour. The reverse side of the bulla is not flat but rather multifaceted. It features a crosswise hole and the impression of a fixing strap, likely made of leather or parchment, measuring 7 mm in width. This strap would have been used to affix the bulla to the object sealed.

Figure 26. Bulla inv. no. 217.

Figure 26. Bulla inv. no. 217.

This bulla bears a single seal impression featuring a well-executed vine scroll depicted with double stroke lines (Figure 27). The scroll forms a spiral, with circles filled with five-lobed leaves. Above the vine scroll is an inscription, while below it is the representation of a mongoose walking to the right. The mongoose has an elongated, slender body with short legs, which contribute to its agile movement. It has a bushy tail as long as its body, a pointed snout, small, rounded ears, and an open mouth with its tongue extended, revealing sharp teeth. Its large, glaring eye adds a dynamic and lifelike quality to the portrayal. The entire composition is encircled within a wreath, enhancing the overall aesthetic of the impression. The vine scroll motif was a popular decorative element during the late Sasanian and early Islamic periods, and this bulla provides an exemplary illustration of its use.

Although there is no mention of the mongoose in Persian literature, Chinese records of the Tang dynasty historical records provide valuable insight into the significance of this animal as a diplomatic gift at the Sasanian royal court. The Jiu Tang shu, 198.5312, which is the classic historical work about the Tang dynasty, records the following account:

In year 21 [of reign-era Zhenguan, i.e., 647], Yazdegerd sent an animal as tribute, called an *ichneumon (mongoose). It was shaped like a rat, but dark in color. Its body length was eight or nine inches. It can enter holes and grab rats. Yazdegerd was cowardly, and being pressed by commanders, he thus fled to Tukhāra, but did not arrive, for he was killed by troops of the Tāzīks [Arabs]. His son, Pērōz, also sought shelter in Tukhāra. Yabghu granted him pardon.

(Kotyak 2024, p. 73)

As Kotyak (2024, p. 73) points out, while it may initially seem unusual for the Sasanian court to send a mongoose3 as a diplomatic gift to the Chinese court, this gesture aligns with the established tradition of offering exotic animals as courteous tokens of diplomacy and goodwill.

The seal impression bears a single word in Middle Persian cursive script, starting anticlockwise from ca. 1 to ca. 11 o’clock. The word reads as follows:

⟨ʾpzwny⟩

/abzōn/

‘increase’

The word /abzōn/ is a common form of blessing that frequently appears on the obverse of Sasanian coins alongside the king’s name. It appears for the first time on coins of Kavād I, from the 16th regnal year, during his second reign (499–531 CE) as ⟨kwʾt ʾpzwny⟩ ‘May Kavād grow’. On a gold dinar where Kavād holds a ring in his hand, the legend reads ⟨lwbʾny ʾpzwn⟩ /ruwān abzōn/ ‘increase for the soul’. Schindel (2004, 3/2, pp. 381, 383–84) translates the phrase as ‘may his soul grow’ or ‘His soul has grown/increased’. The word /abzōn/ is also found on numerous seals and seal impressions from both Sasanian and early Islamic periods.

A second example of the seal impression on Alpaiwalla Bulla 217 appears on a bulla housed in the American Numismatic Society collection (Gropp 1974, p. 136, pl. XXI. A.3). A different, but similar design, accompanied by the word /abzōn/ is also found on a bulla from Qasr-i Abu Nasr in Shiraz (Harper 1973b, p. 81, D42). Here, the design consists of a spiral vine scroll, with one circle filled with a five-lobed leaf and the other a bunch of grapes. The mongoose is absent from that seal impression, but the overall design, style, and execution closely resemble those of the Alpaiwalla 217 seal impression. In these three examples, the stylised leaves bear a close resemblance to late Sasanian and early Islamic motifs seen in metalwork and stucco. The seals that produced these impressions likely date from the late 5th to 7th centuries, a period when cursive scripts were in use (Cereti 2008, p. 188).

7. Bulla inv. no. 218 (Figure 28)

This bulla measures 16 × 17 × 5 mm and weighs 1.54 g. Both the obverse and reverse are reddish brown in colour. The reverse side features a crosswise string hole, designed to allow a string or cord to pass through and bind the object sealed.

Figure 28. Bulla inv. no. 218.

Figure 28. Bulla inv. no. 218.

This bulla bears a single impression of an oval-shaped seal with a flat engraved face measuring 11 × 14 mm. It displays a male portrait bust with the head shown in left profile and the shoulders and chest frontally (Figure 29). His facial features are detailed: a long prominent eyebrow sloping towards the top of the pronounced nose, a large fully opened eye with prominent eyelids, an elongated nose, and full lips. He has a pointed beard. The ear is not shown. The hair is stylised with oblique lines ending in a raised line and then gathered into a scroll framing the forehead and gathered at the back of the head in a ball with lines representing curls. The bust wears a round-collared cloak with a prominent neckline. The drapery is rendered as rows of parallel folds, three of which are visible on the right shoulder. The bust terminates to a curved line and is positioned between two upright wings of conventional form. The outer edges of the wings display traces of feathers, represented by short crosswise strokes. The wings rest on a small base stroke and are likely those of an eagle. This motif is often associated with the yazata Wahrām (Gropp 1974, p. 131; Compareti 2009–2010, p. 38) (NP. Bahrām, Av. Vərəθraγna), the god of war and victory. This assumption is probably based on the interpretation that the word vārəγna-, attested in the Bahrām Yasht (Yt. 14.19), refers to the eagle (Charpentier 1911, p. 59). According to a tradition reflected in Bundahišn 13.22 (Agostini and Thrope 2020, p. 72), the eagle is one of the birds, among many other creatures, created from the ‘seed’ (šusr, Iranian Bundahišn 13.4) of the primordial animal after Ohrmazd had purified it in the moon. However, this association is not conclusive since vārəγna- translates simply as ‘bird of prey’ without specifying a particular species (Bartholomae 1904, pp. 1411–12; Hintze 1994, pp. 198–99).

The style and treatment of the facial features of the bust lean towards naturalism, even though the characteristics of the image are those of the conventional portrait that does not represent a specific individual, such as the bearer of the seal or a sovereign. Instead, it is the depiction of a generalised human male image. Such depictions of male busts are widely represented on Sasanian seals (Bivar 1969, pl. XII). The impression belongs to a personal seal, owned by a private individual who was not necessarily, but could have been, a state official. The absence of typical attributes associated with royal Sasanian iconography such as insignia, fluttering ribbons or kōlāh/kōlāf (headdress) suggests that the individual was not connected to the royal household or held a high-ranking office.

The arrangement of the hair, modelling of the bust, and naturalistic style, as well as the placement of the bust between the two wings of an eagle are also found on a chalcedony seal housed in the Metropolitan Museum of Art, New York (Brunner 1978, p. 57, no. 104) and on several chalcedony seals in the British Museum (Bivar 1969, pp. 45–46, pl. 1. AB, nos. 1–10). Moreover, the treatment of the face, the drapery, and the placement of the bust on a curved line between two wings closely resemble two seals in the Bibliothèque nationale de France (Gignoux 1978, pls. XVI. 5.4, XXIII. 7.8) and a seal in the Louvre Museum (Gignoux 1978, pl. XXVIII. 9.45). However, none of these seals is dated. The seal stone that produced the impression on the bulla from the Alpaiwalla Museum can be dated to the 4th century based on stylistic comparison with those datable seals from the British Museum (Bivar 1969, pp. 45–46, pl. 1. AB, nos. 1–10).

The seal impression bears an inscription in lapidary Middle Persian script circling anticlockwise and starting at 3 o’clock and terminating at 9 o’clock. Our reading of the surviving letters on the circumference is as follows:

⟨ʾps W pn[ʾ]h c[L] lšn⟩

/abes(tān) ud pan[ā]h [ō] rašn/

‘Trust and refuge (in) Rašn’

The characters at the top of the seal impression between 1 and 11 o’clock are severely damaged. Since there are no indications that the upper part of the bulla is broken off, this seems to be due to the way the seal stone was impressed into, or extracted from, the soft clay during the sealing process. Rather than being fragmentary, the first three characters ⟨ʾps⟩ could be the abbreviated spelling of /abestān/ ‘trust’, as is the case, for example, on seal 9.59 and on bulla 17.12a of the Bibliothèque nationale de France (Yamauchi 1993, p. 78; Gignoux 1978, p. 73, no. 9.59, p. 29, no. 9.59, p. 136, no. 12.17a). The word abestān is followed by ud, which is in turn precedes the word ⟨pnʾh⟩, although the letter alef is absent. After ⟨pnʾh⟩, there are five additional letters. The first is illegible, the second appears to be a wāw or šin, and the final three can be confidently read as ⟨lšn⟩. The word abestān ‘trust’ is usually followed by ō and then either by the reference to divine beings (yazdān or mihr) or a concept associated with the sacred fire, such as ādur gušnasp, burz-mihr, or ātaxš ʾpzldwpy. In some cases, abestān is followed by ud panāh, as in abestān ud panāh ō ātaxš gušnasp. When combined with yazdān ‘sacred beings’, it forms the phrase abestān ō yazdān ‘trust in the sacred beings’ (Moradi and Hintze 2026). On our bulla, the formula appears to be an abbreviated form of abestān + ud panāh + [ō] + a divinity (possibly rašn). However, the presence of a wāw or šin before ⟨lšn⟩ complicates this interpretation. Alternatively, the last four letters might be read as warān ‘ram’ or šarān, though both readings are unexpected. One might consider reading the character preceding the lamed of ⟨lšn⟩ /rašn/ as ayin and interpret it as the first character of ⟨ʿL⟩ /ō/ ‘in’, the preposition which usually follows abestān ud panāh. If so, one would have to assume that the engraver has written the character lāmed only once instead of twice.

8. Conclusions

This article has examined six Sasanian bullae from the Framji Dadabhoy Alpaiwalla Museum, contextualising them within the broader sealing practices of the late Sasanian period. Despite uncertainties surrounding their precise provenance, the iconography and inscriptions on these bullae shed light on key aspects of Sasanian administrative system, as well as its cultural and religious symbolism. The seal impressions reflect the intersection of visual culture, bureaucratic function, personal identity, and the integration of religious authority within the administrative apparatus, highlighting the complexity of interpreting sigillographic material where artistic convention and institutional roles converge. These seals thus functioned not only as instruments of authentication but also as vehicles for symbolic communication.

Two of these bullae (nos. 214 and 219) provide important insight into the structure and nomenclature of late Sasanian local administration, particularly the role of the ⟨mgw⟩ /mow/ (priest) and the institution of the mowūh ⟨mgwh⟩. Each iconic and epigraphic impression bearing the ⟨mgw⟩ title belongs to a personal seal, owned by a private individual, who was not necessarily, but could have been, a state official. By contrast, the aniconic impression bearing the term mowūh represents an administrative seal. The iconic seal impression is of the same size as that of the administrative seal. Its size and position on the bulla as well as being both iconic and epigraphic suggest that it is the official seal of the cosignatory. Smaller seal impressions on the edge of the bullae likely belong to the witnesses who were present during the sealing process.

The co-occurrence of the impressions of personal and administrative seals on a single bulla reflects bureaucratic protocol in late Sasanian period, wherein official functionaries and local witnesses authenticated documents and packages jointly. Their presence together on a single legal document highlights the intertwined nature of religious and administrative spheres in the late Sasanian state apparatus. This convergence reflects a bureaucratic system where priestly figures not only held spiritual authority but also participated actively in legal and governmental processes.

Although the term mowūh ⟨mgwh⟩ remains linguistically and functionally ambiguous, its frequent appearance in seal impressions—far outnumbering those of any other administrative designation—attests to its prominence within the bureaucratic system of the late Sasanian and even early Islamic periods. Its exclusive attestation in administrative contexts, and absence from literary and epigraphic sources, suggests that mowūh denoted a localised magian office, likely responsible for civic and possibly judicial duties at the district level. It may have been subordinate to, or operated alongside, the mowbed office. The prominence of this title in sealings such as those discussed here highlights its importance as a key administrative unit active across various provinces of the Sasanian realm.

The imagery on other impressions of personal seals was probably selected not solely for aesthetic purposes, but for its symbolic significance. These images reflect the beliefs, values, and imaginative worlds of the seal holders, shaped by their political, social, and religious contexts.