Twelve Years Without Rain: A Rigvedic Hymn and the Legend That Outgrew It

A King Who Turns Up in Two Books

The Mahābhārata begins its dynastic business with a king called Śantanu: the Kuru ruler who falls in love with the river Gaṅgā, fathers Bhīṣma, and then wants the fisherman’s daughter Satyavatī badly enough to accept a condition that disinherits his first son. The epic’s first book turns on that bargain.

Now go to the far end of the Rigveda’s tenth maṇḍala, into verse composed centuries before anyone thought of fixing the epic in its present form. Hymn 98 names a king. The name is Śantanu. The priest officiating for him is called Devāpi. In the epic, Devāpi is Śantanu’s elder brother, the prince who should have had the throne and did not get it.

Two names, in the right relationship, in two texts separated by most of a millennium. It looks like the one place where the oldest Indian text touches the dynasty of the great epic, and it has been read that way since at least the middle of the first millennium BCE.

Then you read the hymn. There are no brothers in it. There is no throne, no succession dispute, no disinheritance, no curse. What the twelve stanzas record is a piece of business: a king needs rain, a specialist is engaged, the specialist asks the god of speech for words that will work, the water comes, and the fee is counted out in wagon-loads. The hymn is closer to a receipt than to a story.

The distance between those twelve stanzas and everything later tradition said about them is the most interesting thing here. Usually, when a Vedic commentator supplies the story behind a hymn, we cannot tell whether he preserved something old or invented something convenient. In this case the hymn, the earliest story, the intermediate versions and the final form all survive and can be laid side by side. You can watch a legend being built, identify the gap it was built to fill, and read the builders’ reasoning.

Twelve Stanzas, One Transaction

The hymn is in triṣṭubh, the standard eleven-syllable quatrain, and it sits in the stretch of the tenth maṇḍala that Hermann Oldenberg’s metrical and textual analysis treated as the collection’s latest layer, the appendix at the end of the appendix.[13] Its traditional index entry assigns it to Devāpi Ārṣṭiṣeṇa himself: the man inside the hymn is credited with composing it. Such attributions are evidence of what the tradition believed, not of who held the pen, a problem taken up in the piece on the Anukramaṇī’s impossible authors.

What makes the hymn unusual is that it keeps changing who is speaking. It opens as a request, switches into the voice of a god answering, slides into third-person narration about the man who made the request, and ends as instructions to Agni. Griffith’s 1896 translation flattens some of this; the shifts are plain in the Sanskrit.[1]

Verse 2 is the strangest thing in the hymn and the easiest to miss. A god is quoted, in the first person, promising to supply the words with which he will then be addressed:

ā devó dūtó ajiráś cikitvān tvád devāpe abhí mām agachat

pratīcīnaḥ práti mām ā vavṛtsva dadhāmi te dyumatīṃ vā́cam āsán

“The god’s messenger, quick and attentive, has come to me from you, Devāpi. Turn round, face me: I am placing a shining speech in your mouth.”

RV 10.98.2, in the hymn. My rendering from the romanised Saṃhitā, checked against Griffith.[1]

Then the man answers, and what he asks for is not weather.

asmé dhehi dyumatīṃ vā́cam āsán bṛ́haspate anamīvā́m iṣirā́m

yáyā vṛṣṭíṃ śáṃtanave vánāva divó drapsó mádhumām̐ ā́ viveśa

“Place in our mouth a shining speech, Bṛhaspati, free of affliction, quick-moving, by which the two of us may win rain for Śantanu. The honeyed drop from heaven has entered it.”

RV 10.98.3, in the hymn. My rendering; Griffith has “speech lucid, vigorous, and free from weakness”.[1]

Notice vánāva, a first-person dual: “the two of us may win”. Priest and god are a team with a joint deliverable, and the deliverable is rain for a named third party who is paying for it.

Chosen as Purohita

The Rigveda’s very first verse calls Agni purohitá, “the one placed in front”, and that is where the word normally lives: as an epithet of the fire who stands at the head of the rite. Collect every occurrence and the distribution is lopsided.

Methods note. These counts come from a string search of the romanised Saṃhitā, inspected verse by verse against Griffith and checked against Grassmann’s dictionary, still the practical concordance.[17] Forms of purohitá are counted; the abstract purohití is listed separately, and 6.70.4, where the word sits in a locative phrase about electing the Hotṛ, is ambiguous and left out of the tally.

The institution that dominates later Indian political life, the royal chaplain at the king’s shoulder whose rituals underwrite the reign, is in the Rigveda almost entirely a way of talking about fire. The two verses where a named man holds the title are among the collection’s last compositions, and one of them is this:

yád devā́piḥ śáṃtanave puróhito hotrā́ya vṛtáḥ kṛpáyann ádīdhet

devaśrútaṃ vṛṣṭivániṃ rárāṇo bṛ́haspátir vā́cam asmā ayachat

“When Devāpi, placed in front for Śantanu, chosen for the office of Hotṛ, meditated in longing, Bṛhaspati, well pleased, handed him a speech that the gods hear, a winner of rain.”

RV 10.98.7, in the hymn. My rendering, with Griffith’s “chosen for Hotar’s duty” for hotrā́ya vṛtáḥ.[1]

Vṛtáḥ, “chosen”, is the vocabulary of selection, the root that gives the ritual texts their technical term for engaging a priest. Verse 4 tells him to sit down and get on with it. The hymn is documenting an appointment.

And it prices it. Verses 9 to 11 count the offering in adhirátha, “wagon-loads”: a thousand in verse 4, ninety-nine thousand in verse 10, ninety thousand handed over to Indra in verse 11. The word adhirátha occurs in exactly one other place in the Rigveda, the hymn about Mudgala’s strange chariot race, where it is again a unit for counting prizes (RV 10.102.2; see the piece on that race). Nobody delivered ninety-nine thousand wagon-loads of anything. The number is doing what the numbers in Rigvedic gift-praise usually do, which is to make a transfer sound enormous, and the economics behind that habit is the subject of the articles on the poet-patron gift economy and on debt and obligation.

So: a king with a weather problem, a hired professional, a god consulted as a supplier, and an invoice. The Rigveda has plenty of rain hymns, most of them addressed to Parjanya with no patron in sight. It has almost nothing else that reads like a contract.

Speech Without Affliction

Look again at what Devāpi actually asks for in verse 3. Not clouds. Words. The adjective he attaches to them is anamīvā́, “free of amīvā”, where amīvā is bodily affliction, the sickness that charms and healing hymns are written against.

Elsewhere it qualifies what you would expect: wealth and strength (RV 3.16.3), food offerings (3.22.4), worshippers hoping to stay well (3.59.3), the lord of the dwelling who should be “without affliction” towards his residents (7.54.1), the dead man Yama is asked to keep in health (10.14.11), the widows who go up from the pyre “tearless, free of affliction” (10.18.7), the dawns (10.35.6), long life under the sun (10.37.7). Nine verses, and in eight the word is about bodies, houses, herds and days.

Only at 10.98.3 is it about a sentence.

That is a precise and rather beautiful idea: that an utterance can be diseased, that a priest’s formulation might come out crippled, and that what a professional most needs from the god of ritual speech is sound words rather than fluent ones. The Rigvedic poets think about their craft constantly, usually in the vocabulary of carpentry and weaving (see the poet as carpenter and the demand for a new song). Here the metaphor is medical, and it runs through the whole hymn: the last verse asks Agni to drive amīvā away along with the demons, so the same word for sickness appears at both ends, once as a quality the speech must not have and once as an enemy to be expelled.

Aside. A suggestion, offered as a suggestion. The later tradition’s most distinctive addition to this story is that Devāpi was physically afflicted, and most versions specify a skin disease. Nothing in the hymn says so. But the hymn does put the vocabulary of affliction twice in the mouth of this particular speaker, in a collection where that vocabulary is otherwise reserved for bodies. A commentator working on a hymn whose speaker asks for speech “free of affliction”, in a tradition that read etymologically and associatively as a matter of method, had the raw material for an afflicted speaker lying on the page. I would not push this further than a possibility, and there is a reasonable counter-argument: physical defect is a stock disqualifier for kingship in the epic and legal literature, so the motif could have arrived from that direction instead. Either way, both roads run through the same word.

What Yāska Had to Explain

The earliest narrative we have about Devāpi and Śantanu comes from Yāska’s Nirukta, the oldest surviving Indian treatise on etymology and the interpretation of difficult Vedic words, usually placed somewhere around the fifth or fourth century BCE. At 2.10, in Lakshman Sarup’s standard edition and translation, Yāska is working through the hymn, and he stops to supply its circumstances.[4]

His story: Devāpi and Śantanu were brothers, Kurus. Devāpi was the elder. He withdrew to practise austerity, and Śantanu had himself anointed king. From that time no rain fell in Śantanu’s kingdom for twelve years. The brahmins told him why: he had displaced his elder brother. Śantanu went to Devāpi and offered him the sovereignty. Devāpi did not take the throne; he took the office of Hotṛ, sacrificed for his brother, and the rain came.

Every narrative element in that paragraph is new. The hymn supplies a king, a priest, a drought by implication, a ritual and a result. Yāska supplies the kinship, the seniority, the renunciation, the anointing, the twelve years, the diagnosis by priests, the offer of the kingdom and the refusal.

[!NOTE] Yāska is not writing fiction, and precision about what he is doing matters. The Nirukta is a commentary on hard words and obscure passages. When he tells a story, the story is an explanatory device: it answers the question “what situation would make these verses make sense?” Later commentators treated his answers as transmitted history. That is a different claim, and he did not make it.

Here is the same material across the sources, with the additions marked where each one first appears.

Read the right-hand column downwards and watch the problem change. Yāska needs to know why a priest and a king stand together in a drought. The Bṛhaddevatā needs to know why the elder brother did not rule. The epic needs the pair fitted into a genealogy running on to Bhīṣma. The Viṣṇu Purāṇa needs the disqualification to be doctrinal, because by then the live threat to a prince’s legitimacy is heresy rather than dermatology. The latest layer needs him for something else entirely.

The Ākhyāna Problem

This is one case of a question that organised Vedic philology for a century. Many Rigvedic hymns look like fragments: they presuppose a situation they do not describe, change speakers without warning, or begin an action halfway through. Hermann Oldenberg proposed in the 1880s that such hymns are the verse portions of ākhyāna, narratives told in prose with verse at the emotional high points, where the prose was never fixed and therefore never preserved.[9][10] On that reading, the stories told by later commentators might be genuine remnants of the lost prose.

Emil Sieg took the strong version of the position in 1902, trying to recover an Indian itihāsa tradition inside the Rigveda itself and treating the hymns as the surviving verse of real narrative cycles.[11] Devāpi was one of his cases, which is how Śantanu ended up in the footnotes of the debate.

The deflationary reply came from the people who read the hymn very closely. Arthur Macdonell and Arthur Keith, compiling the Vedic Index in 1912, were blunt about this entry: the hymn certainly represents Devāpi as sacrificing for Śantanu, but there is no trace in it of any brotherhood between the two men, and the relationship, with everything that depends on it, enters with the Nirukta and the Bṛhaddevatā.[3] They were equally sceptical about the detail in verse 11, where an obscure word aulāná- is to be set “in heaven among the gods”. Macdonell thought Śantanu himself was probably meant; the word resists confident translation, and it belongs with the collection’s other genuinely untranslatable items. F. E. Pargiter, arguing genealogically, took the patronymic Ārṣṭiṣeṇa as evidence that Devāpi’s father was a king named Ṛṣṭiṣeṇa, which fits the epic’s dynastic frame better than the hymn.[19] John Spellman revisited the tangle in 1959, treating Devāpi’s character and his claim to the throne as problems rather than data.[12]

Oldenberg’s basic observation is sound, and nobody doubts that Rigvedic hymns often presuppose stories their audiences knew; Jamison and Brereton are careful about exactly this, and their guide to the collection says so repeatedly.[2][20] What does not follow is that the commentators had access to those stories. The test is whether a commentator’s narrative explains something in the verses that nothing else explains, or merely fills a hole in a way that suits his own period.

graph TD
    H["RV 10.98: priest, king, rain, fee"] --> N["Nirukta 2.10: brothers, 12 years drought"]
    H --> G["Gap in the hymn: why these two men?"]
    G --> N
    N --> B["Brhaddevata: skin disease bars the elder"]
    B --> E1["Adi Parva: ascetic from boyhood"]
    B --> E2["Udyoga Parva: unfit body, brahmins object"]
    E2 --> P["Vishnu Purana: heretics make him unfit"]
    P --> K["Kalapa village: Devapi never dies"]
    H --> V["Vedic Index 1912: no brotherhood in the text"]
    N --> S["Sieg 1902: lost narrative recovered?"]
    V --> S

By that test the Devāpi legend fails in an informative way. Nothing in the hymn is clarified by making the two men brothers. The verses are perfectly coherent as an employment record, and the brotherhood does not help with a single difficult word, including aulāná-. What the brotherhood does is answer a question the hymn never raises and that a later audience could not help raising: if this priest is a king’s equal in the poem, who is he to the king?

The Brother Who Could Not Be King

Once the brotherhood exists, it needs a reason, and the reason the tradition settles on is a body.

The Bṛhaddevatā, the verse handbook on the deities of the hymns that Macdonell edited and translated for the Harvard Oriental Series in 1904, already has the skin disease.[5] The epic keeps it and sharpens it. In the Udyoga Parva account, Devāpi is the eldest and best loved of King Pratīpa’s sons, but he is afflicted, the brahmins forbid his installation, and the principle is stated flatly: the gods do not accept a king whose body is deficient. Devāpi, described as defective in a limb, sees his father blocked on his account and leaves for the forest.[6] The Ādi Parva’s genealogy, telling the same dynasty’s story, gives a milder version with no disease at all: the elder brother simply became a renunciant while still a boy.[6]

Two explanations inside one epic, for the same man, is a useful reminder of how these texts work. They are not a single author’s account; they are an archive of accounts.

The underlying rule, though, is consistent and it has a long life. Vedic and post-Vedic India expected the king’s body to be unblemished, because the king’s body is an instrument of ritual as much as of government, and a flawed instrument invalidates the rite. The Rigveda itself has a developed vocabulary for bodies that fail and for divine repair, which is the subject of the article on broken bodies and the Aśvins. What the legend does is run that rule backwards into the Rigveda: it takes a priest who asks for speech free of affliction and makes him a prince disqualified by affliction, so that his presence at the ritual rather than on the throne becomes necessary instead of merely contractual.

Kings Who Fail the Sky

There is a second idea the legend brings and the hymn does not have, and it matters more than the skin disease.

In the hymn, the drought is a problem. In Yāska, it is a verdict. Rain stops because the wrong man was anointed and resumes when the error is acknowledged. The weather has become an audit of legitimacy, and the brahmins are the auditors who read the result and tell the king what he did wrong.

That is the political theology that later Indian kingship runs on, and it is not a Rigvedic idea in this form. The Rigvedic version is cruder and, to my reading, more candid. Rain is a commodity; gods have it; poets know how to ask; kings pay. The sovereign in 10.98 is not being judged, he is shopping. Theodore Proferes has traced how Vedic ideas of sovereignty were built out of exactly this material, the poetics of power rather than its ethics, over the later Vedic period.[18] The hymn sits at the start of that process. Yāska sits well along it.

Hold both versions in view and the pair is more interesting than either. One world has a king who can buy his way out of a dry season; the other has a king whose dry season is his own fault. The second is a far more demanding theory of government, and it entered the tradition attached to a hymn that does not contain it.

Names in the Same Pool

So is the Rigvedic Śantanu the Śantanu of the epic? The honest answer involves a short list of other names.

Michael Witzel’s reconstruction of the formation of the Kuru state uses precisely this evidence. On his account the Kurus barely exist in the Rigveda proper: the clan name shows up once, inside the personal name Kuruśravaṇa, and the Kuru material in the Atharvaveda’s Kuntāpa section belongs to a later stage.[14] The Kuru realm then becomes the political centre in which the Vedic canon was collected, ordered and taught, which means the late Rigveda and the epic tradition were assembled in the same region a few centuries apart, by people drawing on the same stock of royal names.[15]

That is enough to explain the coincidence without a dynasty behind it. Two texts rooted in the same country will share names the way two centuries of English kings share Edwards. And the hymn gives us nothing a genealogist could use: no father, no son, no capital, no clan for Śantanu, not even the title rājan. His own hymn does not say he was a king. We infer it because somebody is buying rain in bulk.

One late Rigvedic poem does name a dynasty, and it shows what patron-praise looks like when the poet cares about the man:

“Had I command of the immortal gods, or were I even lord of men, my generous prince would be living still. No one lives beyond the statute of the gods, not with a hundred lives. So I am parted from my friend.”

RV 10.33.8 to 9, after Griffith.[1] The poet, in this hymn, has just named Kuruśravaṇa, his father Mitrātithi and his son Upamaśravas.

Three generations of a Kuru house, in grief, in the same corner of the collection as our hymn, and nobody has ever needed to invent a story to make it legible. That is the contrast: 10.33 says who everybody is, 10.98 does not, and only one of them attracted a legend.

Still at Kalāpa

The last stage still surprises me. By the time the Purāṇas sort out the kings of the Kali age, Devāpi has been promoted out of history. In the Viṣṇu Purāṇa, he and a figure named Maru of the solar line are living still, at a village called Kalāpa, endowed with the power of yoga, waiting.[7] The Bhāgavata Purāṇa states the assignment: at the end of the Kali age the two will return to human society and re-establish the order of social and religious duties.[8] A man who appears in the Rigveda for six verses, as a hired specialist with a weather contract, becomes the tradition’s reserve copy of civilisation.

The hymn’s own last verse, by contrast, keeps its eyes on the clouds:

ágne bā́dhasva ví mṛ́dho ví durgáhā apā́mīvām ápa rákṣāṃsi sedha

asmā́t samudrā́d bṛható divó no ‘pā́ṃ bhūmā́nam úpa naḥ sṛjehá

“Agni, beat back the hostile forces and the hard passages; drive off sickness, drive off the demons. From this great ocean of heaven release for us here the abundance of the waters.”

RV 10.98.12, the hymn’s last words. My rendering; Griffith gives “send down on us a mighty flood of waters”.[1]

No eternity, no dynasty, no vindication. A request for water, and a request that the things which make a season hard should go away.

Reading the Gap

Treat 10.98 as the historical document everyone wanted it to be and it disappoints. It will not confirm the Mahābhārata, date the Kurus, or tell you whether Bhīṣma’s father had an elder brother. The two names sit in the hymn like pins with no thread between them.

Treat it as evidence of something else and it becomes one of the most instructive hymns in the collection. The Rigveda is full of allusions whose stories are gone (see the hymns that read like fragments of lost theatre), and usually we are left guessing whether a commentator preserves or invents. Here the whole chain is legible: the gap in the verses, the first attempt to fill it, its hardening into biography, its adjustment to later anxieties, and its final flowering into eschatology. The tradition could not let a priest and a king stand side by side in a poem without specifying what they were to each other, and once it had decided they were brothers, everything else followed under pressure of plausibility.

Notice also what the oldest layer was comfortable with. A priest who might botch the words. A god who supplies them as stock. A king who pays in wagon-loads for weather. Nobody in the hymn finds any of that embarrassing, and nobody suggests the drought was anyone’s fault.

Read 10.98 straight through, then read Yāska’s paragraph on it in Sarup’s translation, then read the Udyoga Parva’s version. It takes half an hour, and it is the clearest available demonstration of how Vedic India turned poetry into history. For the longer context, there are companion pieces on how the collection was sorted and stratified, on who the poets were, and on the assembly politics of the last hymns.