Panda registers - the bahis
This is the richest genealogical source available to most north Indian families, and the least used.
At the major pilgrimage centres, hereditary priests called pandas (or pandits, tirth purohits) maintain bahis - long cloth-bound registers in which visiting families record their visit, their lineage and the purpose of the pilgrimage. The entries are signed. A family returning a generation later signs beneath the previous entry, in the same book, and the two entries connect.
Some of these registers run back several centuries. They are private property, held by the panda's family and inherited within it.
Where they are kept
- Haridwar - the largest and best known collection, held by pandas at Kankhal and the ghats. Registers here cover a very wide catchment across north India
- Gaya - visited for pitru paksha and shraddha, with correspondingly detailed ancestral information in the entries
- Varanasi - particularly for the eastern plain
- Kurukshetra, Prayagraj, Trimbakeshwar, Nashik, Chintpurni, Badrinath - regional and occasion-specific collections
- Puri - for Odia families
How the system is organised
This is the part that determines whether your visit succeeds. Registers are indexed by place, not by name. Pandas are allocated territory: a particular panda's family serves particular districts, or particular villages, and holds the books for that territory alone.
So the sequence is:
1. Arrive knowing your ancestral village and district - the mool. Without this you cannot be routed to the right panda
2. Ask at the ghat for the panda of that district. The pandas maintain their own allocation and will direct you
3. Give the name of the oldest ancestor you can name, and his father's name if you have it
4. The panda searches his index by village and family, and reads the entry aloud
A successful search commonly yields a written lineage going back five to ten generations, names of relatives who visited, dates of visits, and occasionally the reason for a visit - which can date a death, a marriage or a migration.
Practical notes
Go in person; this cannot be done by phone. Take a written list of every name, spelling variant and village you have. Expect to pay dakshina, agree the amount beforehand, and be prepared to be shown books rather than given copies - photography is sometimes permitted, sometimes not, and the decision is the panda's.
Take a photograph or careful transcription of everything you are shown, on the day. People routinely leave having heard eight generations recited and remembering three.
The Maithil panji
For Maithil families the panji prabandh is the exceptional resource: a formal written genealogical register maintained since roughly the fourteenth century, kept by specialists called panjikars, and consulted before marriages to verify permitted degrees of relationship.
Unlike the pilgrimage bahis, which record visits, the panji is a systematic genealogy - descent, marriages, and the connections between families across generations. Where it covers your family, it can produce a documented line that no other Indian community can routinely match.
Panji material is held by practising panjikars, in family collections, and in institutional archives in Darbhanga and elsewhere. There is no central index. The practical route is to find a panjikar working in your family's area, usually through the family purohit or through relatives who have used one for a marriage within the last generation.
If your family is Maithil and someone in it married in Mithila in living memory, a panjikar was almost certainly consulted at the time, and that consultation is a thread worth pulling.
Land and revenue records
Revenue records are the most reliable dated documents most families can reach, because the state had a durable interest in keeping them.
- Khatauni - the register of landholders and holdings in a village
- Khasra - the plot-level record
- Jamabandi - the record of rights, updated periodically
- Settlement records - from the colonial survey and settlement operations, often the earliest systematic naming of individuals in a village
These name a holder, his father, and frequently the wider co-sharing family. Because the records were revised on a cycle, successive versions show inheritance and partition happening - which is to say, they show the family tree, indirectly but reliably.
Most states now publish current records online - Bhulekh in Uttar Pradesh, Bihar Bhumi, Bhulekh Odisha, and their equivalents elsewhere. Portals and their addresses change often enough that it is not worth listing links here; search for the state name with "bhulekh" or "land records" and use the official state government domain.
Older, pre-digitisation records are held at the tehsil office and in the district record room, and are consulted in person. Ask for the purana abhilekh or the settlement-era khatauni for the village. A patwari or lekhpal who knows the village is worth more than any portal.
Other archives worth knowing
Indenture and emigration records. For families in Mauritius, Fiji, Trinidad, Guyana, Suriname and South Africa, the emigration pass is the key document. It records the emigrant's name, father's name, caste, age, village, district (zilla) and depot of embarkation. Several national archives and university projects have digitised and indexed these; the Mauritius and Fiji collections are the best known. For a diaspora family this document alone can restore a village name lost for four generations.
Temple records. Donation registers, service tenures and mahant records at major temples name families across long periods. Puri, Varanasi, Gaya and the Mithila temples are all worth approaching.
The National Archives of India and the state archives hold settlement reports, district gazetteers, and administrative correspondence. District gazetteers are underrated: they describe the communities of a district as they were at the time of writing, often naming prominent families and their villages.
Census material. The decennial census from 1871 onwards does not survive as individual household schedules in India, so it cannot be searched by name - but the published reports describe distributions and communities usefully.
Army records. Recruitment from Awadh and Bhojpur was heavy. Service records name village and next of kin.
Family paper. Before any archive: check what the family already has. Old wedding invitations, janam patris, court papers, partition deeds, school leaving certificates, passports, ration cards, letters. A single partition deed can settle a generation of guesswork, and there is usually one in a steel almirah somewhere.
Oral history, done once, properly
Everything above depends on having a name and a village to search with, and for most families the only source of those is a living elder.
Do it as a recorded interview rather than a conversation, take ninety minutes, and follow an order - own childhood first, then the household, then the village, then the record-keeping questions, then texture. The full method, with the twenty questions, is set out in Interviewing your grandparents.
Two rules that matter more than the questions: record the audio, and record the consent at the start of the audio. And write the names down in the spelling the speaker used, noting where you are unsure.
Templates
Two things to download. Neither is clever; both are the format that survives.
Vanshavali template (CSV)
One row per person, with columns for names and spellings, approximate dates, village, gotra and - the important one - the source of each fact. Opens in Excel or Google Sheets.
Download →Interview sheet (text)
Twenty questions in the order that works, with space for answers. Print it, or keep it open while recording.
Download →How to record what you find
The discipline that makes a family record last is citing your sources - not formally, just honestly.
For every fact, note where it came from: the person and the date you spoke to them, the register and the panda's name, the document and where it is kept. Mark approximations as approximations. Record uncertainty explicitly: "possibly two brothers, informant unsure" is more valuable than a clean tree that quietly guesses.
Keep two copies in two houses. Digital files in one person's laptop are not a record; they are a single point of failure with a battery.
And when you have something, consider contributing it. Family histories and vanshavalis submitted here are reviewed and, with the consent of the living people named, published as a page under the family's own name - the contact form is the way in. Nothing is published without consent, and anything published can be removed on request.
यह वेबसाइट पारिवारिक एवं सांस्कृतिक विरासत का दस्तावेज़ है। स्रोत भिन्न हैं और मौखिक परंपराएँ बदलती हैं - सुधार सादर आमंत्रित हैं। सुधार भेजें.