The second vīthī of the northern (Airāvata) mārga. TEXTUAL PROBLEM: Matsya 124.55 as transmitted reads nāgavīthī for this road, repeating the previous name — certainly corrupt for Gajavīthī. Both witnesses checked (GRETIL's Chaukhamba text and sa.wikisource) share the error, so it is old; the 1916 translator flags it, noting the text Wilson quoted should read Gajavīthī. Wilson's own note 21 to VP 2.8 lists Gajavíthi among the nine.
गजवीथी Gajavīthī
the Elephant-road · — · Taurus/Orion (Rohiṇī, Ārdrā, Mṛgaśiras)
Attestations
Transcribed from the Devanagari OCR of the printed edition at shloka_source_url (Achyuta Nand Jha ed., vol. 1) and checked word-for-word against two further independent scans of two different printed editions of the same commentary — Krishna Chandra Dvivedi's (https://archive.org/download/brihatsamhitawithbhattotpalavivrittipartidr.krishnachandradvivedi/Brihat%20Samhita%20%20with%20Bhattotpala%20Vivritti%20Part%20I%20-%20Dr.%20Krishna%20Chandra%20Dvivedi_djvu.txt) and A.V. Tripathi's (https://archive.org/download/JiUm_brihat-samhita-part-1-of-varaha-mihira-by-avadh-bihari-tripathi-saraswati-bhavan-granthamala/Brihat%20Samhita%20Part%201%20of%20Varaha%20Mihira%20by%20Avadh%20Bihari%20Tripathi%20-%20Saraswati%20Bhavan%20Granthamala_djvu.txt). Only glyph-level OCR artifacts that all three scans show to be artifacts (e.g. OCR रद्वव for रद्गव, श्व for श्च) have been normalised; every substantive reading is as scanned, and the two places where the scans genuinely differ are flagged in identification_notes.
FIVE INDEPENDENT SCHEMES, four of them pre-Varāhamihira, are set side by side by Utpala in ch. 9 — this is the single densest surviving witness to the vīthī system, and the database's vīthī entries hitherto rest on the Purāṇas alone. (a) DEVALA and others: nine triads straight from Aśvinī, Nāgā to Dahanā. (b) KĀŚYAPA: the same triads, given asterism by asterism. (c) VARĀHAMIHIRA'S OWN (BS 9.2–3), which starts the Nāgā with Svātī instead of Aśvinī and gives the Go-vīthī four asterisms and the Dahanā two. (d) PARĀŚARA, differing again (Airāvatī four, Vṛṣabhā two, Mṛga-vīthī two, Ājī = Anurādhā/Viśākhā/Jyeṣṭhā, Vaiśvānarī = Mūla + the two Āṣāḍhās). (e) GARGA, close to Varāhamihira. Plus (f) the SAMĀSA-SAṂHITĀ, Varāhamihira's own lost epitome, agreeing with (c). Utpala explicitly notes the clash — 'atra varāhamihireṇa saha bhedaḥ', 'here there is a difference with Varāhamihira' — and Varāhamihira himself refuses to adjudicate at 9.7. The Gaja-vīthī is the ONE vīthī on which every authority Utpala quotes agrees: Rohiṇī, Mṛgaśiras, Ārdrā. The database's entry for it rests on a conjectural emendation of Matsya 124.55; Utpala's five witnesses put the content beyond doubt.