There is a spot on the southern bank of the Godāvarī, where the Peṅgaṅgā runs into it, that is said to end rebirth outright for anyone who bathes there and gives alms.2
It has four names. Vṛṣākapi. Hanumat. Mārjāra. Abjaka.
Rai Bahadur Lala Sītārām reports a story from the Brahma Purāṇa that he says explains all four at once, and the claim is worth taking seriously for a moment before taking it apart, because a single narrative that accounts for four separate names looks like strong evidence. Four coincidences resolved by one cause is exactly the shape a real explanation has.
What it turns out to be is two different stories that happen to have washed up at the same riverbank — and an inference drawn from that fact which the essay makes in passing and never examines.
This page is about the inference, not the tīrtha.
The tīrtha story starts with the vanara king Kesarī, who has two wives.
Both were apsaras once, and both are under a curse that took their faces. Añjanā's face became a vanarī's. Adrikā's became a cat's.2
They live on Mount Añjan, and the sage Agastya visits. They worship him well; he is pleased; he grants each of them a son. Vāyu's union with Añjanā produces Hanuman. Nirṛti's union with Adrikā produces Adri, who becomes king of the piśācas.2
Then the curses come off — bathing in the Godāvarī removes both. And here the story does the thing it was built to do. The place where Adri bathed Añjanā becomes the tīrtha called Añjan and Paiśāch. The place where Hanuman bathed Adrikā becomes the tīrtha called Mārjāra, Hanumat and Vṛṣākapi.2
Notice the courtesy in the arrangement. Each son washes the other's mother. Nobody bathes his own.
Read the naming and it is almost entirely mechanical. Mārjāra means cat, and Adrikā's cursed face was a cat's. Hanumat is the son who did the bathing. Añjan is the mother, Paiśāch the son who bathed her. Five of the six names are a straight record of who was standing in the water and what they looked like.
This is what pilgrimage etiology does. A place accumulates names, the names need accounting for, and a story is generated that puts somebody at each one. Causation runs from the name to the story, not from the event to the name — which is why these accounts are always so tidy and always so crowded.
Vṛṣākapi is the exception, and it is the exception because nothing in this story explains it. No cat, no bath, no mother. The name is simply also attached to the site, and the tīrtha narrative hands it to Hanuman along with the two it has actually earned.
So the essay goes and finds where Vṛṣākapi comes from, and it comes from chapter 129 of the same Purāṇa, and it has nothing to do with monkeys.
Hiraṇya's son Mahāśani grows invincible by his own power, binds Indra, and presents the king of the gods to his father as a trophy. Then he attacks Varuṇa.3
Varuṇa does not fight. He makes peace by giving Mahāśani his own daughter. With Indra still captive the gods go to Viṣṇu, who tells them he can do nothing without Varuṇa's cooperation — a striking sentence, and one the essay passes over.3
So the gods go to Varuṇa, who talks Mahāśani into releasing Indra. But not before Indra is humiliated and instructed to treat Varuṇa as his guru from then on. Indra goes home ashamed and tells Indrāṇī.3
Indrāṇī's advice is austerities to Śiva and Viṣṇu in the Daṇḍaka forest. The two of them perform the penance together at the Godāvarī–Peṅgaṅgā confluence, and it produces a being carrying both gods' forms at once — trident and discus together — who goes down to the netherworld and kills Mahāśani.3
That being, Indra's own beloved friend, is called Abjaka Vṛṣākapi.
Here is the essay's move, in full: the same tīrtha is tied to both stories, therefore Vṛṣākapi and Hanuman were one and the same figure.3
That is the whole argument, and it is worth being precise about why it fails, because the failure is common and it is not the failure of being religious.
A shared location is evidence about the location. Two narratives attach to one confluence. That establishes the confluence was important enough to attract narratives — which we knew, since it promises the end of rebirth. Popular sites accumulate stories the way a bulletin board accumulates notices, and two notices on one board are not two accounts of one event.
The essay would need something the place cannot supply: a text naming them together, a non-generic shared attribute, a lineage, an episode in common. What it has is that both stories end at the same riverbank — the one thing every story told at a tīrtha shares with every other.
And the sequence runs the wrong way. The tīrtha story does not derive Vṛṣākapi from anything; it inherits the name and assigns it. The chapter-129 legend explains where the name came from. Taken in order, chapter 129 accounts for the name and the tīrtha story then borrows it — which is the opposite of two independent attestations converging.
Sītārām does not notice. The identification arrives as a subordinate clause, not as a thesis, and nothing later in the essay depends on defending it. That is the most telling detail here: it is not a bad argument he made and lost. It is an argument he did not know he was making.
The same essay carries two accounts of Añjanā's curse and they disagree about who cast it.
In the Vālmīki material — Kiṣkindhā 66, where Jāmbavān recites Hanuman's parentage to a Hanuman who has been sitting silent through the leap-auction — Añjanā was the apsara Puñjikasthalā, cursed into vanara form by the sage Atri.1 That attribution is unusual; the curse is more commonly laid elsewhere, and this build records it as a variant rather than smoothing it.
In the Brahma Purāṇa tīrtha account the curse has no author at all. Both wives are simply "under a curse," and the sage who appears in the story — Agastya — is there to give rather than to take, granting the sons.2
Sītārām prints both and reconciles neither, which by this volume's standards is normal practice rather than negligence. But it compounds the Vṛṣākapi problem. An essay that will not adjudicate between two curse-attributions in adjacent paragraphs is not an essay whose identifications should be leaned on.
Sort the two accounts by what each is for and they stop competing.
The Vālmīki version is about a person: it exists to explain how Hanuman came to be capable of the leap, and it sits inside a scene where somebody needs persuading of exactly that.
The Brahma Purāṇa version is about a place. It tells a pilgrim why the water they are standing in ends rebirth, and it needs a cast large enough to cover every name carved at the site. Hanuman is in it because he was available and locally famous, not because it is about him.
That is the finding worth keeping. The corpus's habit of reaching one destination by several unrelated routes has been recorded on this build before, as a sign that the destination was fixed and the routes were reasoning. Here the same mechanism runs in reverse: a place was fixed, and a figure was recruited to it. Hanuman does not acquire a tīrtha because of anything he did. He acquires it because a tīrtha needed somebody, and he was the name in the district.
Use this as the test for any identification offered on shared-context grounds — both of these are associated with X, therefore they are the same. It appears constantly in devotional literature, in popular etymology, in brand history, in genealogy, and wherever two records happen to name one place, date or institution. Three questions, in order.
One: what does the shared thing attract? If the context is popular — a pilgrimage site, a famous city, a widely used platform — the coincidence has a base rate near one and carries no information. A tīrtha promising liberation is the least rare context available.
Two: which direction does the derivation run? Establish which account explains the shared element and which merely inherits it. Chapter 129 explains Vṛṣākapi; the tīrtha story receives it. An inherited element is not a second attestation, it is the same attestation seen twice.
Three: what would the claim predict elsewhere? A genuine identity should show up off the shared site — a joint appearance, a non-generic shared epithet, a lineage. A claim that predicts nothing outside its one coincidence is not an identification but a filing convenience.
The status of the material. Everything here is one essay in English rendering, reporting two portions of the Brahma Purāṇa at second hand.23 The chapter number for the Vṛṣākapi legend (129) is given; the tīrtha material's chapter is not. [PLAUSIBLE — needs corroboration] on the narrative details; [LOW CONFIDENCE] on the identification, which is the essay's own inference and is labelled as such above.
The absence claim is bounded. Nothing here says the Brahma Purāṇa nowhere identifies Vṛṣākapi with Hanuman. It says this essay's argument for it rests on the shared tīrtha and nothing else, which is a claim about the essay and is checkable against the essay. The Vedic Vṛṣākapi of Ṛgveda 10.86 is a genuinely contested figure with its own long scholarly literature; whether any of that literature supports the identification is not addressed here and was not read.
Open question. Does any text name Vṛṣākapi and Hanuman together outside a tīrtha context? One such passage would convert this from a coincidence into a lead. Its absence would settle the matter the other way.
⚠ Limitation, recorded and closed. This essay's closing section is an extended piece of colonial-era racial philology. It is excluded from this vault by ruling of 2026-08-14 and is recorded once, as a fact about the 1975 volume's contents, in the Limitations section of the source stub. It supplies no etymology, no footnote and no handshake here, and the legitimate portions above are written from the Vālmīki and Brahma Purāṇa material only.
Sītārām is a careful reporter and an incautious reasoner, and the two habits sit in the same paragraphs.
The reporting is genuinely good. He gives the Kiṣkindhā sarga number, names Puñjikasthalā, keeps the Atri attribution though it is not the common one, and lays out the tīrtha's full cast without tidying the awkward parts. Nothing in his summary of what the sources say has to be taken on trust.
The reasoning goes soft only where the conclusion is one he has no stake in. He does not argue for the Vṛṣākapi identification; he arrives at it, which is what people do with conclusions they think are free.
Against the volume's better close-readers that is instructive. Vedprakāś Śāstrī's epithet method can fail against the actual order of a verse, which is why it is citable. Sītārām's tīrtha inference could not have come out the other way — and that, not devotion or scholarship, is the difference between them.
The Tīrtha Launders Anything. That page has a Purāṇic tīrtha working as a costed instrument: a cow establishes that the expiation is available, prices a murder against it, and proceeds — the location entering the decision in advance. The two pages show one kind of site doing two jobs, neither of them narrative. There a tīrtha absorbs a sin nobody could otherwise carry; here it absorbs a name nobody could otherwise account for. The pairing yields a rule about sacred sites as institutions: a tīrtha is a general-purpose sink, and what gets discharged into it is whatever the surrounding literature cannot otherwise dispose of. Which predicts that the stronger a site's claim, the more heterogeneous the material accumulating there — and the less any one accumulation tells you about the others.
Nīlakaṇṭha's Vedic Hanuman. Both are retrospective identifications of Hanuman with an older figure, and they fail in opposite directions. Nīlakaṇṭha attribute-matches, reading Hanuman out of two Ṛgvedic verses by lining up three qualities, one of which is a task rather than a quality and carries the whole weight. Sītārām context-matches and never looks at attributes at all. The diagnostic: attribute-matching over-fits, context-matching under-determines, and a reader motivated toward a conclusion will use whichever is locally available. Neither man chose a method. Nīlakaṇṭha had verses and no place; Sītārām had a place and no verses. The identification was the constant.
The uncomfortable part is that the reasoning this page criticises is the reasoning everybody uses, and it usually works. Two documents naming the same obscure village probably are related. Two people at the same tiny conference plausibly know each other. The inference is sound and cheap, which is exactly why it fails invisibly — the moment the shared context becomes popular the base rate goes to one, and nothing about running the inference feels any different from inside.
Sītārām did not feel sloppy. He felt like somebody noticing a connection, which is what noticing a connection feels like whether or not there is one.