[disclaimer: I tend to use Indic-style spellings in most cases]
One of the questions often pondered upon is the state of the Iranian religion before the ascendancy of the cult of zarathuShTra in greater Iran. There are multiple facets to this question, some of which include: 1) The state in the “heartland” where zarathuShTra and his band came into being. 2) The state of the religion of the early Aryan occupants Iran, i.e. the of the pre-Achaemenid Medians and Persians. 3) The nature of religion among the non-Persian Iranians like the shAka haumavarga, shAka tigrakhauda, shAka taradArya and kushAnas. 4) The history of iconic worship amongst Iranians.
The answers are largely murky, but some facets are clear. The cult of zarathuShTra clearly came from within a pan-Indo-Iranian system that was quite close to the ancestral version of the shrauta system of the Aryans of India. It is also clear that zarathuShTra-s elevation of the asura varuNa prototype in his system was not a pure innovation. Even in the Vedic system we observe that asura varuNa held an exalted position often combined with the great indra in the devatA dvandva indrAvaruNa of tremendous power. We also notice that within the Vedic system certain vipra-s mention a personal connection with the great asura varuNa, like vasiShTha and bhR^igu-s/atharvan-s. From the fragments of the pre-Achaemenid Medians and Persians, and also to certain extant the later Parthians and shAkas we observe that the system based on the worship of the Aditya-s (though they are never called that) was the dominant one. From names we discern the prominent worship of mithra: mithra is seen in personal names like mithrapAta, mithrabarzana or dAtamithra or mithradAta. A “shrine” of mithra is mentioned as being maintained by the Iranians in Fayum, Egypt at the time of Alexander’s invasion, and a certain mithradAta is mentioned in its connection. There was the feast of mithra celebrated through the Achaemenid times, the mithrakAna festival. The term ahura just by itself or as mAzdAdAta or mAzdAfarnah (in Aramaic) or the term mAzdAyazna are all encountered. However, strikingly, bhaga is also prominently mentioned, and perhaps most frequent in pre-Achaemenid names. There is the holy place named bhagasthAna, i.e. modern Behistun. Another place name with bhaga is bhagadAta, i.e Baghdad. There is a month named bhAgayAdi. There are dvandva constructions like bhaga-mithra and bhaga-mazda and also forms like mithra-bhaga and mithra-ahurA-mAzdA. We notice that unlike other vipra families that have long pre-R^igveda pedigrees in the Vedic world, the brothers vasiShTha, agastya and the lesser known mana suddenly appear claiming divine origin from mitra and varuNa and the patronymic maitravaruNi. Of the vasiShTha is one of the most prolific composers and has several hymns to varuNa or mitrAvaruNa and also interestingly supplies the longest solo hymn of bhaga in the entire shruti. It is quite possible that these maitravaruNi-s came into the Indic realm from the borderland with Iranians where the Aditya-based system was already rising in dominance.
The worship of indra was always around in the midst of the Iranians and persisted under the specific name of verethraghna after zarathuShTra’s demonization of the name andra. It is possible that it was prominent in the non-Persian Iranian groups like the shAka haumavarga as suggested by the fragmentary third party accounts of their worshiping a certain “war-god” who in all likelihood was the indra archetype. However, in much of the Iranian zone, unlike in the early Indic zone the worship of indra-archetype was probably secondary to the mitra-varuNa archetypes. Likewise, the mighty vAyu, even after zarathuShTra’s heterodox sweep, remained the first receiver of the haoma in the Iranian ritual. From the Kushana evidence we note that vAyu was extremely prominent and perhaps the preeminent deity in the eastern Iranian frontier (also note an echo of this in the vedic soma offering and the animal of vAyu – vAyavyam pashu, where vAyu receives the primary offering). Even after zarathuShTra demonized the ashvin-archetype under the name nanghaitya (Vedic: nAsatyA), he recycled them as the twin amesha-spentas: haurvatAt and amR^itAt. We have earlier alluded to the possibility that the pANDava-s of the mahAbhArata were actually from the Iranian borderlands. In fact their origin mythology retains this Iranian ordering: yuddhiShThira: dharma (the archetype of mitra and varuNa, the upholders of dharma; ahura mAzdha and mithra, upholders of the Asha=R^ita); bhIma (vAyu); arjuna (indra, verethraghna, Herakles in the Greek world); nakula-sahadeva (the ashvin-s, haurvatAt and amR^itAt). Note indra is only third in the system as in the Iranian system.
This issue of the marginalization of certain deities to the central Aditya, the varuNa-archetype, is also reflected in the emergence of public iconic worship amongst the Iranians. As we have discussed before the Iranians, like their Indic counterparts, were largely un-iconic in public ritual to start with, probably using icons only in small private shrines. In fact the cult of zarathuShTra was a vigorously iconoclastic one which appears to have been imbibed by the Judaistic stream of the Semitic world. But on the other hand other Middle Eastern Semites and Elamites appear to have influenced the Iranians towards iconic worship. However, the most prominent forms of public iconic worship emerged via the Greek-Iranian contact. The famous edifice at Nemrud Dagi built by Antiochus son of Mithridates claiming descent from both Alexander and Darius is a splendid example of this. The gigantic images of deities depicted there were: ahura mAzdha equated with Zeus; mithra equated with Apollon; and verethraghna with Herakles or Ares. Thus even though we know that Zeus is actually the indra-archetype, here he has been equated ahura mAzdha who has eaten into indra’s preeminent position. Instead the actual ortholog of indra, verethraghna has been reduced to an equation to the “junior Zeus” Herakles or just the “war-god” Ares.
On the eastern front things were dramatically different, with Iranians receiving iconographic inputs not just from the Greeks but also Indians. For example vAyu of the Kushanas inherits his iconography from rudra in India. This was further strengthened by the homophony of their names: vAyush>oesho and Isha. Just as Zeus was equated with ahura mAzdha we find that among the Kushanas ahura mAzdha was depicted as indra riding airAvata. Again the Kushana verethraghna gets depicted as similar to kumAra just as he was equated with Ares in the Nemrud Dagi images. The goddess aredvI sUrA anAhitA, the sarasvatI ortholog, was one of the Iranian deities to receive an iconic representation from very early on. It appears that even Artaxerxes-II, who had a special devotion to this goddess, might have founded iconic temples of hers at Hamadan, Babylon and Damascus. In the early period her iconography was influenced by Semitic goddesses such as Ishtar. But in the eastern zone of the Iranian territory her iconography was clearly influenced by chaNdikA or durgA. Like sarasvatI (also perhaps even the prototype of Pallas Athena, sarasvatI and anAhitA), she had a warrior aspect. Both Sassan and Babak appear to have been priests of a temple of anAhitA at Istakhr before the latter went on to found a new Iranian empire. She was routinely invoked by his successor Ardeshir for victory in battle. Thus, it is not surprising that in the eastern region her depictions came to closely resemble the Indian warrior goddesses.
Several interesting iconographic developments that remain poorly understood because of the disaster of Afghanistan appear to have happened at the Indo-Iranian border zone. For example there is deity named “mana-bhaga” who appears to have been depicted like a hybrid of saMkarShaNa and vAsudeva with a chakra and a halAyudha. The exact nature of this “Iranian pA~charAtra” remains entirely unclear. But somewhat later pA~ncharAtra icons of nArAyaNa with an Indo-Iranian syncretism are also known from Afghanistan. For example, a smuggled image from 427 CE, sold to private collector in the US, has the inscription that was translated by Doris Srinivasan as: “In the year three, in the month of aShaDha, on the fifth day, on this day the image of nArAyaNa was installed at bhImAsthana in gharaTTamaTTha of the shrI-vaihlika-s. It is a donation of… shrI varisha.” The image is a typical vaiShNava of the chaturmukha viShNu type, though only 3 faces are actually visible, flanked by the images of sudarshana and gadA-devI. The side heads are the usual nR^isiMha and varAha heads, but the central head is quite distinctive. This head has a clearly Iranian crown and moustache. Thus an Iranian strain of syncretic vaiShnava imagery appears to have been reasonably prevalent in gandhara and bAhlika. Interestingly, the earliest proto-types of the trishirobhairava or later dattAtreya imagery appear to have also come up at least by the 200s of CE in the same region – the three-headed form with the signs of shiva, viShNu and brahmA.