A custom is law only when it is proved to be law. Unlike a statute, of which a court takes judicial notice, a custom set up in derogation of the general law must be established as a fact — by evidence, in the particular case, by the party who asserts it. The burden of proof of custom is therefore one of the most practically important rules in the whole field of customary law: it decides who must lead evidence, how strong that evidence must be, and what happens when the proof falls short. This article traces the rule from the Privy Council in the Ramnad case and Hurpurshad v. Sheo Dyal through to the Supreme Court in Saraswathi Ammal and Ujagar Singh, and explains the statutory machinery in Sections 13 and 57 of the Evidence Act that governs how custom is pleaded and proved.
Custom as a Fact to be Proved, Not a Law to be Noticed
The starting premise is that a custom, when relied upon to displace the ordinary text law, is treated by the courts as a question of fact. A statute carries its own authority and a court is bound to take judicial notice of it; a custom carries no such inherent authority and must be brought home to the court by proof in each case. The classic statement is that of the Privy Council in Hurpurshad v. Sheo Dyal (1876) 3 IA 259, where it was observed that a custom is "a rule which, in a particular family, or in a particular district, has from long usage obtained the force of law," and that it "must be ancient, certain and reasonable." Because the custom derives its force from usage rather than from the general law, the person who sets it up must show that the usage has in fact existed and has acquired the force of law.
This characterisation — custom as fact — has two consequences that run through every case in this area. First, the ordinary rules on burden of proof under the Evidence Act apply: he who asserts must prove. Second, until a custom has been so frequently proved in litigation that it becomes notorious, the court will not assume it; it must be demonstrated afresh. As we shall see, this fact-based treatment shapes both the incidence of the burden and the standard to which it must be discharged. The point connects closely with the definition and nature of custom, where the same ancient-certain-reasonable triad recurs.
The Ramnad Case: Clear Proof of Usage Outweighs the Written Text
The foundational pronouncement on the strength of proven custom is the Privy Council's decision in the Collector of Madura v. Moottoo Ramalinga Sathupathy (1868) 12 MIA 397, universally known as the Ramnad case. There the Board declared that "under the Hindu system of law, clear proof of usage will outweigh the written text of the law." The case concerned the validity of an adoption made by the widow of the last Zamindar of Ramnad, and the Privy Council, after surveying the smritis and commentaries, upheld the adoption on the strength of established usage in the Dravida country permitting a widow to adopt with the assent of her husband's sapindas.
The significance of the Ramnad case for the burden of proof is twofold. On the one hand it elevates custom to a position of primacy: a clearly proved usage will prevail even over an apparently contrary smriti text, which is why custom is so powerful a source of Hindu law. On the other hand, that primacy is conditioned entirely on the word "clear." The usage prevails only when it is clearly proved; a vaguely asserted or loosely evidenced practice carries no such force. The whole of the burden-of-proof jurisprudence is, in a sense, an elaboration of what the Privy Council meant by the single adjective "clear."
On Whom the Burden Lies
The cardinal rule is that the burden of proving a custom lies on the party who asserts it. Where a litigant seeks to displace the general law — whether the textual Hindu law, the rules of Muslim law, or a statutory rule of succession — by setting up a custom to the contrary, it is that litigant who must establish the custom. The Privy Council in Hurpurshad v. Sheo Dyal put it plainly: where members of a family set up a custom derogatory to the ordinary law, the burden lies on them to prove it. The court does not begin with any presumption in favour of the alleged custom; it begins, if anything, with the ordinary law and requires the asserting party to dislodge it.
This allocation is simply an application of the general principle in the Evidence Act that the burden of proof lies on the person who would fail if no evidence were given on either side, and that he who asserts the affirmative of an issue must prove it. A custom is an affirmative, special fact relied upon to take the case out of the general rule, so the asserting party must prove it. The corollary is equally important: if the party setting up the custom leads insufficient evidence, the custom fails and the general law applies by default. The party resisting the custom need prove nothing affirmative; he succeeds simply on the failure of the other side to discharge its burden.
The Standard: Clear, Cogent and Unambiguous Evidence
It is not enough to lead some evidence of a practice; the evidence must be clear, cogent and unambiguous. The custom must be proved to be ancient, certain and reasonable, and the usage must be shown to have been acted upon with such invariability and for so long a period that it has, by common consent, been accepted as the governing rule of the particular family, caste, community or district. The requirement of invariability is critical: a practice that has been followed sometimes one way and sometimes another cannot be a custom, because it lacks the certainty that gives custom the force of law. These standards mirror the essentials of a valid custom, since proof and validity are two sides of the same coin — the court asks whether what has been proved satisfies the legal requirements of a binding custom.
The Privy Council in Mst. Subhani v. Nawab AIR 1941 PC 21 gave the standard its most quoted formulation. The Board held that the law requires that a custom, however ancient, must in the absence of clear proof of its continuous observance and invariability fail to obtain legal recognition; a custom "however old it may be, in the absence of clear proof of its continuous observance without any variation, would not obtain legal existence." The case concerned whether, under the customary law of the Tulla clan in the Punjab, collaterals could take precedence over married daughters in non-ancestral estate, and the proof was tested against this exacting standard. The lesson is that antiquity alone is never enough; continuity and invariability must be proved with equal care.
Antiquity: Indian Law Does Not Insist on Immemorial Origin
English law fixed a rigid test of antiquity: a custom had to be shown to have existed since "time immemorial," technically from the year 1189 (the first year of the reign of Richard I), failing which it was not a valid custom. Indian courts have consistently refused to import this artificial limit. In Hurpurshad v. Sheo Dyal, the Privy Council expressly held that it is not essential in every case to carry the antiquity of the usage back to a period beyond the memory of man; it depends on the circumstances of each case what degree of antiquity must be established before a custom can be accepted.
This relaxation was reaffirmed in Mst. Subhani v. Nawab, where the Privy Council again declined to insist that the custom be immemorial, holding instead that the antiquity required is a matter of degree to be judged on the facts. The practical effect is that an Indian litigant need not prove an unbroken chain back to a fixed historical date, but must prove an origin sufficiently ancient that the practice can be said to have hardened into law, coupled with continuity and invariability. The custom cannot be of recent or modern origin; a usage that began within living memory and has not yet acquired the settled character of law will not qualify. The point is developed further in the discussion of the essentials of a valid custom.
Section 13 of the Evidence Act: Transactions and Instances
The mode of proving a custom is governed in part by Section 13 of the Indian Evidence Act, 1872 (now mirrored in the Bharatiya Sakshya Adhiniyam, 2023), which provides that where the question is as to the existence of any right or custom, the following facts are relevant: (a) any transaction by which the right or custom in question was created, claimed, modified, recognised, asserted or denied, or which was inconsistent with its existence; and (b) particular instances in which the right or custom was claimed, recognised or exercised, or in which its exercise was disputed, asserted or departed from.
Section 13 is thus the statutory engine for proving custom by evidence of instances. A party seeking to establish a custom typically tenders specific instances — documented adoptions, mutations, partitions, family settlements, prior judgments, revenue records or wajib-ul-arz entries — in which the alleged custom was acted upon. Each such instance is a relevant fact under Section 13. The greater the number of consistent instances, and the more they span different families and generations, the stronger the inference that an invariable usage exists. Conversely, instances in which the custom was departed from are equally relevant and may negative the alleged custom, since variability defeats certainty. The illustration to the section — concerning the right to a fishery proved by deeds, mortgages and instances of exercise — captures the method precisely.
Instances Must Be Proved, Not Merely Asserted
Because custom is established by instances, the quality of those instances is decisive. The courts have repeatedly cautioned that mere oral assertions of belief in a custom, unsupported by concrete instances of its actual exercise, are worth little. A witness who deposes only that "it is our custom" without pointing to occasions on which the custom was in fact followed adds nothing of evidential value. What carries weight are proved instances — ideally documented and spread over time — in which the practice was observed.
At the same time, the courts have warned against the opposite error of demanding the impossible. In Ahmad Khan v. Channi Bibi (discussed by the Privy Council in the line of cases on Punjab custom and cited with approval in Saraswathi Ammal v. Jagadambal), it was held that a subordinate judge erred in putting aside a large body of evidence merely because specific instances had not been proved. The correct course is to weigh the whole body of evidence — general repute, opinion of persons likely to know, entries in records of rights, and such instances as exist — and to ask whether, taken together, it establishes an invariable usage. Thus the law steers between two extremes: bare assertion will not do, but the absence of textbook-perfect instances will not automatically defeat an otherwise convincing body of proof.
Saraswathi Ammal v. Jagadambal: When Proof Falls Short
The consequences of failing to discharge the burden are starkly illustrated by the Supreme Court in Saraswathi Ammal v. Jagadambal AIR 1953 SC 201. The dispute arose in the dasi (dancing-girl) community of the Tanjore district. Thangathammal, a dasi, died leaving three daughters, and the contest was whether, by a custom of the community, a dasi daughter was a preferential heir to her dasi mother in derogation of the ordinary law of succession. The party setting up the custom led evidence to establish it.
The Supreme Court rejected the alleged custom as not proved. It reiterated that a custom must be ancient, certain and reasonable, and that the burden of establishing it lay on the party asserting it. On examining the evidence, the Court found it inadequate and unsatisfactory: the instances relied upon were too few and too equivocal to establish an invariable usage of the requisite antiquity and certainty. Because the custom failed for want of clear proof, the ordinary law applied. Saraswathi Ammal is the standard modern authority for the propositions that the burden lies on the asserting party, that the standard is exacting, and that an unproven custom simply collapses into the general law.
Ujagar Singh v. Mst. Jeo: No Presumption of a General Custom
The Supreme Court returned to the burden of proof of custom in Ujagar Singh v. Mst. Jeo AIR 1959 SC 1041, a Punjab succession dispute between a deceased landowner's sister and his agnatic collateral. The question was whether, by general custom in the Punjab, collaterals took precedence over a sister in the succession to non-ancestral property. The appellant collateral relied on a supposed general custom; the respondent sister resisted it.
The Court refused to treat Punjab custom as a single homogeneous body of law of which it could take notice. It observed that each tribe in the Punjab has its own customs, and that the existence of a general custom must itself be proved by satisfactory evidence rather than presumed. Reviewing a large number of conflicting decisions, the Court held that a general custom entitling collaterals to succeed in preference to a sister had not been established, and decided in favour of the sister. Ujagar Singh thus underscores two points: there is no presumption in favour of an alleged general custom, and the burden of proving it remains squarely on the party who relies on it, even where the custom is said to be widespread. This is consistent with the treatment of kinds of custom — local, general, family and class, each of which must be separately proved according to its character.
Family Custom: An Even Stricter Standard of Proof
A custom peculiar to a single family is, by its nature, harder to prove than a local or class custom, because it rests on the practice of one family alone and the instances available are necessarily fewer. The courts have accordingly required family customs to be established by strict and conclusive proof. The Privy Council in Mahomed Baksh v. Hosseini Bibi (1888) 15 IA 81 (ILR 15 Cal 684) laid down that a family custom set up to vary the ordinary law must be proved by satisfactory evidence, though the court will not insist on the rigorous and technical rules of proof that English law would apply to such a claim.
The balance struck is therefore a demanding but realistic one. On the one hand, because a family custom is exceptional and is asserted by interested parties, the court scrutinises the evidence with care and requires that the custom be shown to be ancient, invariable and continuous within the family. On the other hand, the court does not demand a quantum or formality of proof that would make a family custom impossible to establish. The asserting party must show a clear and consistent course of dealing within the family over time; isolated instances or self-serving assertions will not suffice. The same logic governs class and caste customs, where the practice of the relevant group must be shown to be uniform.
Once Proved: The Burden Shifts to Show Discontinuance
The burden of proof is not static. Once a custom has been duly proved, or has become so well established by repeated judicial recognition that the court takes notice of it, the burden shifts. The party who then alleges that the custom has been abandoned or discontinued must prove that discontinuance. This principle flows directly from Hurpurshad v. Sheo Dyal, where the Privy Council held that when a custom has once been established, the burden of proving its discontinuance lies on the party who asserts that it has ceased to exist.
The shift is logical. A proved custom has the force of law within its sphere, and law is presumed to continue until shown to have ended. Just as the original burden lay on the party seeking to displace the general law by custom, so the subsequent burden lies on the party seeking to displace the established custom by alleging its extinction. The party asserting discontinuance must show, again by clear evidence, that the practice has been abandoned or has fallen into such disuse and variation that it can no longer be regarded as the governing rule. Mere instances of non-observance may not suffice if the custom is otherwise shown to be alive; the discontinuance must be as clearly proved as the custom itself originally had to be.
From Proof to Judicial Notice: Section 57 and Notoriety
A custom that begins life as a fact to be proved may, through repeated establishment in the courts, ripen into one of which the court takes judicial notice, so that it need no longer be proved in each fresh case. The mechanism lies in Section 57 of the Evidence Act, under which a court may take judicial notice of certain facts, and in the broader principle of notoriety. The leading exposition is the Privy Council's in Effuah Amissah v. Effuah Krabah [1936] AC 83, where the Board held that customs must in the first instance be proved by calling witnesses acquainted with them, until the particular customs have, by frequent proof in the courts, become so notorious that the courts take judicial notice of them.
The Privy Council added a practical suggestion of enduring value: that courts should, in suitable cases, rule which customs they consider proper subjects of judicial notice, specifying the tribes or districts concerned, and ensure that such rulings are reported in an accessible form, so that the body of judicially-noticed custom grows in an orderly way. The same trajectory is visible in Indian law: certain customs — for example, particular usages of succession or adoption repeatedly upheld for a given community — have become so well settled that they no longer require fresh proof. But the threshold is high, and until a custom attains this notoriety the ordinary burden of proof governs. The interaction of proof and judicial notice is a recurring theme in the study of custom as a source of Muslim law as well, where local usages must likewise be proved before they will be applied.
Reasonableness, Certainty and the Limits of Proof
Even a custom that is clearly proved as a matter of fact will not be enforced if it fails the legal tests of validity. The burden of proof therefore operates alongside, not instead of, the substantive requirements that the custom be certain, reasonable, not opposed to public policy, and not in derogation of any statute. A proved usage that is uncertain in its scope, or that offends morality or public policy, or that is expressly abrogated by legislation, will be rejected however well the instances are established. Thus the asserting party must prove not only that the practice exists but that what is proved amounts in law to a valid custom.
This is why proof and validity are intertwined. The court, in weighing whether the burden has been discharged, is simultaneously asking whether the practice shown is ancient enough, certain enough, invariable enough and reasonable enough to constitute a custom recognised by law. A custom that is repugnant to justice, equity or good conscience, or that has been superseded by codifying statutes such as the Hindu Succession Act, will not be saved by however many instances are tendered. The burden of proof is the gateway; the essentials of validity are the test applied once the gateway is passed. For the full architecture of these requirements, see the essentials of a valid custom and the broader introduction to customary law.
Frequently asked questions
Who bears the burden of proving a custom?
The party who asserts the custom bears the burden. Where a litigant sets up a custom to displace the general law, he must prove it. The Privy Council in Hurpurshad v. Sheo Dyal held that where members of a family set up a custom derogatory to the ordinary law, the burden lies on them to establish it. There is no presumption in favour of an alleged custom; if it is not proved, the general law applies.
What is the standard of proof for establishing a custom?
The standard is clear, cogent and unambiguous evidence. The custom must be shown to be ancient, certain, reasonable and invariably observed. In Mst. Subhani v. Nawab AIR 1941 PC 21 the Privy Council held that a custom, however old, will not obtain legal existence in the absence of clear proof of its continuous observance without variation. Variability defeats the certainty that gives custom the force of law.
Must a custom in India be immemorial like in English law?
No. English law required a custom to date from 1189 (time immemorial), but Indian courts have rejected this rigid test. In Hurpurshad v. Sheo Dyal and Mst. Subhani v. Nawab it was held that antiquity is a matter of degree depending on the facts, and the usage need not be carried back beyond the memory of man. It must, however, be sufficiently ancient and not of recent origin.
How is a custom proved under the Evidence Act?
Section 13 of the Indian Evidence Act, 1872 makes relevant any transaction by which the custom was created, recognised, asserted or denied, and particular instances in which it was claimed, exercised or departed from. A custom is thus typically proved by tendering consistent instances — adoptions, mutations, family settlements, revenue records and prior judgments — showing an invariable usage. Instances of departure are equally relevant and may disprove the custom.
What happens if a custom is not adequately proved?
It fails, and the ordinary law applies. In Saraswathi Ammal v. Jagadambal AIR 1953 SC 201 the Supreme Court rejected an alleged custom of the dasi community that a dasi daughter was a preferential heir, finding the evidence inadequate and unsatisfactory. Because the custom was not clearly proved, the general law of succession governed. An unproven custom simply collapses into the general law.
Once a custom is proved, who must prove that it has ended?
The party alleging discontinuance. Per Hurpurshad v. Sheo Dyal, once a custom is established the burden shifts to whoever asserts that it has been abandoned. Because a proved custom has the force of law and law is presumed to continue, the party claiming its extinction must prove the discontinuance by clear evidence — mere isolated instances of non-observance may not suffice.