since this was a downloadable Word document, I decided to give y'all a hyperlink. This here is just a straight copy-paste job. I left in all the spelling mistakes and typos because I reckon they add to the charm. There are squares interjected into a few words since apparently the stolen version of Microsoft Word my buddy put on my computer doesn't have all the goofy alphabet characters installed/activated

A note on the ‘Saeculum’ and the Kushan Era


Robert Bracey


Harry Falk (2001) has recently published a new interpretation of the formula of Sphujiddhvaja for calculating the date of a previously unknown ‘Kushan’ era from the Saka era. A well-known theory, proposed by Lohuizen (1949), that the Kanishka era had continued in a second century, with the hundreds omitted, gave the opportunity to link this new Kushan era with the problem of the date of Kanishka. If Sphujiddhuaja was referring to the start of the second century of Kanishka’s era then this would fix the date of Kanishka to 127 AD.

Though the theory won general support strong objections were raised in correspondence at the time[i], and in more detail once the article was published (Bracey, 2004). The objections to the dating of 127 AD for the Kanishka era were two-fold. Firstly, that the new evidence failed to invalidate the very extensive previous evidence deployed for a variety of other dates, principally 115 AD and 144 AD (Cribb, 1999; Puri, 1994). This evidence is quite extensive and had not been considered, prior to 2001, to favour a date of 127 AD. It therefore remains necessary to explain the discrepancy between the new theory and the old evidence. The second objection stems from the new source itself.

That the Kushan inscriptions represented two sequences of dates is widely accepted, one from the years 2 to 98 of the Kings Kanishka, Huvishka, and Vasudeva; and a second sequence known to include inscriptions dated from the years 20 to 41 and to incorporate the reigns of Kanishka II, Vasishka, and Kanishka III[ii].  It is not clear how these two sequences related to each other. Lohuizen (1949 & 1983) argued that the second sequence was a second century of the era of Kanishka, with the character for 100 omitted. This was plausible, as abbreviation is a common feature of Kushan period inscriptions. But, an alternative opinion (Rosenfield, 1968; Cribb, 1994) was that the second sequence might be a new era. This made sense as the second sequence appeared to commence with the ascension of Kanishka II, and while the foundation of eras was common in the period the abbreviation of dates was unattested.

The calculation employed by Sphujiddvaja (Kushan = Saka + 149), means that his Kushan era starts counting in AD 227. This is hard to equate with the abbreviation of the ‘dropped hundreds’, which is commonly compared to the modern practice of writing ‘05’ or ‘98’. A modern writer never becomes confused over whether ‘05’ means ‘2005’ or ‘1905’, but this is exactly what seems to have happened to Sphujiddvaja. If the inscriptions had been abbreviated then a contemporary author would have understood that the era began in AD 127 and would havewritten the formula accordingly. It would not have been appropriate to be a hundred years out as it would produce nonsense applied to an event in the reign of Vasudeva, or the later Kushan king Shaka. So far from being a neat fit with Lohuizen’s ‘dropped hundreds’ theory Sphujiddvaja actually implied that she was wrong and that a new era had commenced with the second sequence of inscriptions in 227 AD. A valuable contribution but not a successful solution to the date of Kanishka.

In response Falk (2004) has revised his theory with arguments, and new evidence, intended to counter these criticisms. Firstly, he has abandoned the idea of an abbreviation and replaced it with a cyclical era, as well as supporting evidence that cyclical eras existed in contemporary India. Secondly, he has supplied evidence of the continuation of the era, with the hundreds missing, at a still later date. Several aspects of his paper require comment.


1. The ‘Saeculum’


Falk has made explicit the position that has been implicit in his recent argumentation. He argues that far from being an abbreviation (the ‘dropped hundreds’) the era was cyclical by its very nature. This is necessary to explain the discrepancy between Sphujiddvaja and Falk’s theory[iii], but also to explain why amongst hundreds of private donations we do not find a single example of an unabbreviated date[iv]. If the idea of a cyclical era was firmly embedded with the people of Northwest India, then this would explain why not one of the hundreds of private donors ever had a date inscribed with an ‘unabbreviated’ date.

The central case (Falk, 2004: 168-9) is that the concept of a saeculum existed in contemporary ancient India. Falk believes, for reasons that are not clear, that the term would appear as varsaśatam in Sanskrit, and is able to find three examples of equivalent Prakrit terms, in inscriptions of the Iksavakus and Sātāvahanas. He says ‘An allusion to 100 years seems most probable’ (169), and in fact that is exactly the conclusion drawn by Mirashi (1967) who believed that the ‘special meaning in addition to the pure registration of the date’ (Falk, 2004: 169) was a benediction that the king’s life (or rule) should endure for one hundred years. However, Falk is not making this argument, he equates the term with saeculum and so implies that it represents a cyclical era of 100 years. It is therefore implicit that the Iksavakus are employing a saeculum and that since they are contemporary with the Kushans this strengthens the case for a Kushan saeculum.

While not expressing a view on the complicated problem of Iksavaku chronology Table 1 gives an outline of dates available for various Kings from the Eastern Indian site of Nagarjunikonda. This includes two of the inscriptions mentioned by Falk as being dated in a saeculum. Even the most cursory glance at the table shows that these are regnal years and not some sort of cyclical era. If the dates which carry the term vasasatāya were as Falk suggests references to a saeculum then Iksavaku rule would extend over a period of several centuries, and individual reigns would have had to endure for more than fifty years. Any workable chronology of these rulers (who are directly related to each other) would become quite impossible.

The same problems apply to the Sātāvahana mention of vasasatāya. So whatever this word means in the three inscriptions Falk has highlighted it obviously does not mean saeculum. It seems more likely, that it is a general epithet, as Mirashi assumed, which has no special relevance to the reckoning.




Table 1: The Iksavaku Dynasty in Inscriptions



Inscription No.








911, 912, 914-919




Vogel’s M3 and containing the phrase vasasatāya



Exactly the same context as 891 but no mention of vasasatāya








Ehuvala Chantamulasa


886 and 887



681 and 682













Containing the phrase vasasatāya and first believed by Sircar to be date in the 60 year Jupiter cycle.









Reign of Ehuvala Chantamulasa?




Successor to Ehuvala Chantamulasa?

Vasusena, the Abhira

2 or 30


Son of Vasithi?


2. The ‘saeculum’ elsewhere in contemporary documents


There is, however, an example of a saeculum type dating in a well-known Indian source. The various puranas which provide lists of the dynasties of the Kali age also provide a brief account of astronomical calculations (Partiger, 1913: 75). This is translated by Partiger as follows:


“In the circle of the lunar constellations, wherein the Great Bear revolves, and which contains 27 constellations in its circumference, the Great Bear remains 100 years in (i.e. conjoined with) each in turn.”


Often known as the era of the Seven Sages, the term they employ is not varsaśatam, but śatamsamāh (and also saptarsi). Though the term is not that which Falk expects the reckoning described is exactly the sort of dating system that is needed, as it clearly involves a cyclic period of one hundred years.

Dating the puranic texts is immensely difficult because they are heavily stratified. Not only were sections written at different times but they have been subject to subsequent editing. The text undoubtedly contains layers which are drawn from the Kushan/Gupta period. In particular the Puranic texts dealing with the Kali age dynasties provide detailed lists of Kings up to the early Guptas, though they relegate the Greeks, Indo-Parthians, Kushans, and others in the north-west to simple notes on the length of their dynasties. If the astronomical information belonged to the same layer as the historical data it would seem to have been current in the fourth or fifth century and in eastern and central India. 

So, in conclusion, it can be said that while Falk’s argument for varsaśatam being an Indian equivalent of the Roman saeculum is clearly flawed, there may exist a contemporary Indian equivalent.


3. The Dual-Dated Gupta Records


In the second part of his argument, demonstrating the continuation of the era, Falk (2004) draws attention to three inscriptions which are all dated in two different reckonings. In two of the cases both the dates survive, and in one there is a lacunae in which he presumes a date occurs. Falk argues that two epithets, vijayarājya and kālānuvarttamāna, identify two distinct reckonings[v]. If this is the case then a higher date (insc. 515) proceeds a lower date (insc. 933 & 634), a sure sign of a cyclical era.

It is impossible to decide if kālānuvarttamāna represents a specific era, as it occurs only in these inscriptions, so there is no basis for comparison. Instead the assertion that vijayarājya means the Gupta era can be checked. If the term is specific to the Gupta era then the argument holds, but if the epithet can be used with any era then the three inscriptions can be reconstructed without a higher date proceeding a lower (which was in fact how they were interpreted before the present claims).

The term is also employed in the inscriptions of Kumaragupta (insc. 525 & 530) and Skandagupta (insc. 538)[vi]. These are clearly in the Gupta era. A plate of Indravarman dated in the year 87 of the Ganga era (insc. 845) also uses the term. This makes it clear that the epithet does not refer specifically to the Gupta era. And the Omgodu Grant of Sihavarman II in the year 4 (insc. 939) and a grant of Kumara Vishnu dated in the year 13 (insc. 584) use the term as well. Both are clearly in the reign of kings, and this demonstrates clearly that vijayarājya has no special meaning and could be applied as an epithet to any reckoning. It could therefore be applied to either of the reckonings in the dual-dated records.

The claim that two terms (vijayarājya and vasasatāya) have special meanings for the reckoning they accompany has now been demonstrated to be without foundation. It would therefore be unreasonable to claim that kālānuvarttamāna must refer to a specific era, which means that the low dates found in the dual-dated inscriptions could refer to a reign and the high dated inscriptions to the Gupta era; the reversal of epithets, which is the centre of the argument for these being a cyclical era, being irrelevant[vii].

Falk’s argumentation is unsound, but it also ignores a much more compelling line of reasoning first identified by Iyer (1973:21). The low numbered reckoning begins in 426 AD, or 107 of the Gupta era What reckoning could begin in this year? The last inscription of Chandragupta is year 93 (insc. 518) and the first of Kumaragupta is year 98 (insc. 525) so 107 cannot be the start of the reign of Kumaragupta. Iyer recognised this and proposed that the reckoning was the reign of a local king Nripamitra (insc. 704). While Nripamitra is a local potentate and his reign begins close to this date it is a somewhat awkward solution as neither of the dual dated inscriptions mentions Nripamitra[viii].

So while Falk’s argument for a cyclical era being proven directly from the dual dated inscriptions must be rejected (as vijayarājya does not as he claims refer to the Gupta era), these inscriptions do clearly provide evidence of a reckoning that begins in 426 AD. Before considering what era could be reckoned from that point it is necessary to re-examine something which has received only slight attention from Kushan scholars.


4. The Laukika Era


The one example of a cyclical era in Indian history is the medieval Laukika (Salamon, 1998: 196). By way of introduction it is worth quoting what the Arab author Al Biruni says about this era:


“Common people in India date by the years of a centennium, which they call samvatsara. If a centennium is finished, they drop it, and simply begin to date by a new one. This era is called lokakâla, i.e. the era of the nation at large. But of this era people give such totally different accounts that I have no means of making out the truth. In a similar manner they also differ among themselves regarding the beginning of the year” (O’Brien, 1997: 20)


No serious link between Al-Biruni’s Laukika era and the reckoning of Kanishka has previously been drawn, other than as an example to show the plausibility of a ‘dropped hundreds’. The reasons are obvious. The eras are separated by nearly a millennia; the cyclical Laukika era, in which year 1 coincides with 24 or 25 of our own centuries (X24/5) was not close to any of the proposed dates for Kanishka (78, 115, 128, 144 AD); and the ‘dropped hundreds’ was used only to explain a short continuation into a second century, not a lengthy continuation over a long period. However, opinions on both the Laukika and Kanishka eras have changed.

The first change, is that Harry Falk’s new argumentation is now considering a cyclical era as a serious proposition for the Kushan era. The Puranic evidence mentioned above as providing a contemporary example of a cyclical era is connected in later tradition to the Laukika era (O’Brien, 60-3), though it is clear (because the astronomy involved is nonsense) that the Purana’s are providing a post-hoc explanation of an already existing era (Sastri & Sarma, 1959). And the date Falk has proposed, 127 AD, is close to X24/5[ix].

The second change is that the Kushan era has moved closer to the Laukika, as there clearly is an era beginning in 227 AD, and if the dual-dated Gupta inscriptions are to be linked with the Kanishka era (which is obviously open to considerable doubt) they provide evidence for the continuation of the reckoning up to 426 AD. On this presumption then the Kanishka era comes two hundred years closer to the Laukika.

The third change is that the Laukika era has moved closer to the Kushan era. The Bactrian documents of central Asia (Simms-Williams, 1997) provide dates in the Laukika era from the eighth century, and inscriptions from Kashmir and Sind also provide examples of the era considerable predating the chroniclers Al-Biruni and Kalhana. A coin minted in the seventh century (year 88 = 712 AD) may even provide the earliest example (O’Brien, 1997). So the gap between the latest Kushan date and the earliest Laukika has fallen from ten centuries to just three.

The fourth change is the possibility that a cyclical era can lose time. The Laukika era has been demonstrated to lose a year in the twelfth century (O’Brien, 1997), when it conforms to the local reckoning in the Sindh region, and the start of the year is shifted several months backwards, this changes it from being equivalent to AD X25/6 to being equivalent to AD X24/5. A cyclical era is vulnerable to this sort of loss in time, either due to gradual erosion caused by changes in the calendar, or a switch from current to elapsed reckoning, or difficulties of equating specific era dates and AD dates, or simply as the result of the sort of confusion that Al-Biruni complains about.

These four changes affect the relevance of the Laukika era for the Kushan. To illustrate the point Table 2 shows various known reckonings, from the second to the twelfth centuries. Those after the seventh century are known to be cyclical, where those before are only suspected. If it is assumed for the sake of argument that the dates shown represent a single era, then the table shows a gradual loss of time. Like a watch which is running just a little bit slow the era loses time (about 3 months each century). The reckoning begins by ticking the centennium out at X26/7, then X25/6 and finishes in X24/5.



Table 2: Examples of reckonings from the second to the twelfth centuries

Kanishka’s Date







Kushan Era







Dual Dated Gupta Inscriptions







Conquest of Sind Coin







Laukika Era in Bactrian Documents







Laukika Era in Kashmiri Inscriptions







Al-Biruni & Kalhana’s Laukika Era







Jaisalmeri Laukika







Sindh Era

624/5 -626/7






Later tradition of the Seven Sages








The possibility must be entertained that the Kushan era mentioned by Sphujjidvaja is the Laukika era.  Or if that option is rejected, it must be seriously considered that the dual-dated Gupta inscriptions relate to the Laukika era (with which they coincide precisely) and not to the Kushan era.[x] A reckoning which appears to commence in 426 AD and which might be cyclical shows an obvious affinity to the X26 Laukika era of the Bactrian documents. But it also shows some affinity to the Kushan era of 227 AD. So it is unclear if the dual-dated inscription is the latest example of a Kushan date, the earliest example of a Laukika, or both.


5. The starting point of the X27 era


Given then that there is a local reckoning at Mathura, which begins in 426 AD, testified by the inscriptions above, and a local reckoning that begins in 227 AD, testified by Sphjiddhvaja, and if it is assumed (for arguments sake) that they are part of a cyclical era, when does it begin?

What is needed is some clear evidence that the start of the reign of some king coincided with year 1, (with the inauguration of Kanishka I in 127 AD, or with Kanishka II’s inauguration in 227 AD). If neither king initiated the X27 reckoning then it is extremely unlikely that it would coincide with their inaugurations (p ~ 0.094). Should either, or both, kings inauguration coincide with year 1 of a cyclical reckoning it would be almost certain they founded it. If Kanishka II came to the throne in 227 AD and his reign coincided precisely with year 1 of the Kushan era, the conclusion would be forced that Kanishka II was the founder of the era, distinct from the era founded by Kanishka.

The evidence for Kanishka I is compelling. The earliest inscriptions which can be firmly linked to his reign are dated in the year 2 (insc. 181). There are no inscriptions in older eras mentioning king Kanishka and in the Rabatak inscription (insc. 567) he is credited with having ‘inaugurated year 1’. While the evidence for his having begun an era is therefore compelling it is unclear if that reckoning coincides with 127 AD (the balance of the evidence favouring a date from 110 to 130 AD).

The evidence for Kanishka II is also suggestive. It is clear from coin finds in Central Asia that because Kanishka II’s are the last coins to be found in the region that he was the ruler of the Kushan empire who lost Bactria to the Sasanian empire (Zeymal, 1997). That event probably occurred in 233 AD, the foundation of the Kushano-Sassanian era (Sims-Williams, 1997). This puts the start of his reign, suspiciously, close to 227 AD. Some authors have suggested that Vasudeva was still ruling in the first four years of the new century, and that Kanishka II did not come to the throne until year 5[xi]. This is based principally on an assumption that the king Po-t’iao, who sent an envoy to China in AD 230, is Vasudeva (Zurcher, 1968: 371). Though there are no inscriptions of the years 1 to 4, which name Vasudeva as Kushan king[xii].

So, if the idea of an X27 era is accepted, it remains an open question whether that era should be linked to Kanishka, or his successor Kanishka II, or perhaps to somebody else entirely.


6. Conclusion


            Harry Falk deserves considerable credit for advancing Kushan studies towards a solution of the date of Kanishka. In 2000 those strongly interested could pin Kanishka’s date to a period of about thirty years in the first half of the second century, but the trail that had brought that progress since 1960 had gone cold. New lines of evidence, such as the Rabatak inscription or Kushano-Sassanian era had proved somewhat disappointing in this regard. However Falk has twice been able to draw attention to previously neglected evidence (2001 & 2004) and this has given a fresh impetus to solutions, as well as forcing some much needed re-examination of assumptions.

            It has been shown Falk’s evidence of a cyclical era in vasasatāya is unsound, his attempt to equate vijayarājya with the Gupta era is weak, and that therefore the arguments he has recently brought forward cannot be considered compelling. Though it has also been shown that there are alternative lines of evidence both for a cyclical era and which can be used to extend the evidence for a cyclical era, and the problem of how to determine its initial year has been considered. However, a cyclical era still rests on linking a series of different reckoning (Kanishka’s reckoning beginning 110-130 AD, Sphujiddvaja’s Kushan era, the reckoning of 426 AD, and the X25/6 Laukika era). Compelling as they appear these links remain unproven, and this is the central point of disagreement with Falk. Falk has provided some challenging new evidence, but it is not a solution to the problem. At the moment it is only a programme for research.




All inscriptions are referred to by an index number. References can be found by consulting the list of Kushan related inscriptions at





Bhandarkar, D R

1981 Inscriptions of the Early Gupta Kings, CII Vol.3, ASI

Bivar, A D H

1979    ‘The Azes Era and the Indravarman Casket’ South Asian Archaeology 1979: 369-376

Bracey, R

1994    ‘The Date of Kanishka’ A Rough Guide to Kushan History,

1995    ‘Review: Kusana Coins and History’ A Rough Guide to Kushan History,

Cribb, J

1997    ‘Numismatic Perspectives on Chronology in the Crossroads of Asia’ Gandharan Art in Context (ed. Allchin R., Ancient India & Iran Trust: 215-30

1999    ‘The early Kushan kings: new evidence for Chronology. Evidence from the Rabatak Inscription of Kanishka I’ Coins Art and Chronology (ed. Alram M & Klimburg-Salter D E): 177-206

Falk, H

2001    ‘The Yuga of Sphujiddhuaja and the era of the Kusanas’ Silk Road Art and Archaeology 7: 121 – 136

2004    ‘The Kanishka Era in Gupta Records’ Silk Road Art and Archaeology 10: 167-176

Gupta, P L & Kulashreshtha, S

1994    Kusāna Coins and History, D K Printworld, Delhi

Harmatta, J.

2001    ‘Religions in the Kushan Empire’ in History of Civilizations of Central Asia (ed. Harmatta J.), Vol. II, UNESCO

Iyer, S S

1973    ‘Two Brahmi Inscriptions from Mathura’ Epigraphia Indica 40 (1973-74): 19-22

Lohuizen-de-Leeuw, J E van

1949    The Scythian Period: An Approach to the History, Art, Epigraphy and Palaeography of North India from the 1st Century BC to the 3rd Century AD

1986    ‘The Second Century of the Kanishka Era’ South Asian Studies 2: 1-9

Mirashi, V V

1967    ‘Epigraphical Note’ Epigraphia Indica 37 (1967-8): 70-3

O’Brien, A G

1996    The Ancient Chronology of Thar: The Bhāttika, Laukika, and Sindh eras. Oxford University Press, Delhi

Puri, B N

1994    ‘The Kushans’ History of Civilizations of Central Asia (ed. Harmatta J.), Vol. II, UNESCO: 247-264

Rosenfield, J M

1967    Dynastic Art of the Kushans, University of California Press

1968    ‘The Mathura School of Sculpture, two contributions to the study of Kushan chronology’ Papers on the Date of Kanishka (ed. Basham A L): 247-258

Salamon, R

1998    Indian Epigraphy: A Guide to the Study of Inscriptions in Sanskrit, Prakrit, and other Indo-Aryan Languages, Oxford University Press

Sastri, T  S K & Sarma K V

1959    ‘The Untenability of the Postulated Saka of 550 BC’ Journal of Indian History 39.2: 201-224

Shrava, S.

1993    Dated Kushana  Inscriptions, Pranava Prakashan.

Simms-Williams, N

1997    New Light on Ancient Afghanistan: The Decipherment of Bactrian, SOAS

Zeymal, E V

1997 'Coins from the excavations of Takhti-i-Sangin' Studies in Silk Road Coins and Culture: Special Issue of Silk Road Art and Archaeology




[ii] Dissenting opinions include those of Shrava (1993) and Harmatta (1996). These opinions depend upon theories of dual kingship and are unable to explain in an adequate manner the later coinage of the Kushan empire.

[iii] Of course if it were accepted that the two sequences of Kushan inscriptions were separate eras the problems disappear. Though this is not by any means a complete solution.

[iv] The one example, an inscription of year 170 in the reign of Vasudeva (insc. 326) has been misread.

[v] Many epithets in dating formulas are not specific and could be applied to any era. However there are several examples of epithets which are era specific. Several Kharoshti inscriptions use the name of the Indo-Parthian Azes to identify an era commence in the mid first century BC (Bivar, 1979), and the Vikrama era inscriptions from Central India are identified first by the term ‘krita’ and later by the term ‘malwa’ (Bhandarkar, 1981: 187-201). A recently discovered inscription (insc. 23) with dates in three reckonings also uses not only the name of Azes, but also the term ‘yona’ to denote the Greek era (insc. 23).

[vi] These are not from the Mathura area, being respectively from Bilsad, Karamdamda, and Indor, and do not include a second date in a different reckoning.

[vii] It is unfortunate that no example has been found with two high dates. Though it would be dangerous to draw any conclusions from this, as reconstructions of Kushan chronology inevitably involve curious coincidences.

[viii] Inscription 936 may be relevant. This belongs to roughly the same period and has a date in the year 13. Likewise inscription 598 which may read saptarsaya could be relevant to the problem at hand.

[ix] And it is closer still to the Laukika era of X26. In the Bactrian documents alongside the Kushano-Sasanian and Arab reckonings there is an example of the Laukika era in which year 32 = 857 AD, which means that year 1 = X26 (Simms-Williams, 1997: 9)

[x] One of the similarities between the Laukika and Kanishka’s reckoning is their popular nature. Though inscriptions are often referred to as ‘Kushan’ only 10% can be categorised as royal or official. The vast majority of donors that use the era are private individuals who have no connection with the Kushan rulers. This is very similar to the Laukika era which Al-Biruni describes as the popular era of Afghanistan, Pakistan, Kashmir and Sind (almost exactly the extent of the Kushan empire). By contrast other eras are far more commonly associated with royal/official inscriptions; 22% of Azes & Greek era inscriptions, 72% of Gupta era inscriptions, and 87% of inscriptions dated in the reign of a king (based on a sample of 927 inscriptions, May 2005).

[xi] Attempts to separate the inscriptions of Kanishka I and Kanishka II between the years 2 and 19 have been an almost unmitigated disaster. Lohuizen (1949) and Rosenfield (1967 & 1968) have attempted to do so on stylistic grounds, but both failed to correctly place the inscriptions of Vasishka and Kanishka III, there methods indicating they belonged to the first sequence. While attempts to use titles (Gupta & Kulashreshta, 1994) can be shown to be badly flawed and involve considerable special pleading (Bracey, 2005). There are in fact only a handful of inscriptions which can be placed in the second sequence with any confidence (insc. 210, 469, 227, 226, 231, 232, 456) and these provide little evidence on the reign of Kanishka II other than it ending in year 19.

[xii] Joe Cribb has recently presented a series of die studies, which have a direct bearing on this, though the details are as yet unpublished. See